University of Tennessee, Knoxville University of Tennessee, Knoxville
TRACE: Tennessee Research and Creative TRACE: Tennessee Research and Creative
Exchange Exchange
Doctoral Dissertations Graduate School
5-2017
Influence of Patient Satisfaction of Total Knee Replacement Influence of Patient Satisfaction of Total Knee Replacement
Patients on Stair Negotiation and Walking Biomechanics, Patients on Stair Negotiation and Walking Biomechanics,
Strength, and Balance Strength, and Balance
Kevin Alan Valenzuela University of Tennessee, Knoxville, [email protected]
Follow this and additional works at: https://trace.tennessee.edu/utk_graddiss
Part of the Sports Sciences Commons
Recommended Citation Recommended Citation Valenzuela, Kevin Alan, "Influence of Patient Satisfaction of Total Knee Replacement Patients on Stair Negotiation and Walking Biomechanics, Strength, and Balance. " PhD diss., University of Tennessee, 2017. https://trace.tennessee.edu/utk_graddiss/4504
This Dissertation is brought to you for free and open access by the Graduate School at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Doctoral Dissertations by an authorized administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact [email protected].
To the Graduate Council:
I am submitting herewith a dissertation written by Kevin Alan Valenzuela entitled "Influence of
Patient Satisfaction of Total Knee Replacement Patients on Stair Negotiation and Walking
Biomechanics, Strength, and Balance." I have examined the final electronic copy of this
dissertation for form and content and recommend that it be accepted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy, with a major in Kinesiology and Sport
Studies.
Songning Zhang, Major Professor
We have read this dissertation and recommend its acceptance:
Joshua T. Weinhandl, Rebecca A. Zakrajsek, Jeffrey Reinbolt
Accepted for the Council:
Dixie L. Thompson
Vice Provost and Dean of the Graduate School
(Original signatures are on file with official student records.)
Influence of Patient Satisfaction of Total Knee Replacement Patients on Stair Negotiation and
Walking Biomechanics, Strength, and Balance
A Dissertation Presented for the
Doctor of Philosophy
Degree
The University of Tennessee, Knoxville
Kevin Alan Valenzuela
May 2017
ii
Copyright © by Kevin Alan Valenzuela
All rights reserved
iii
DEDICATION
This dissertation is dedicated to my family and friends. I want to thank my parents, John
and Penny, who have always encouraged me from the time I was a child to go for it, regardless
of what it is. I want to thank Denise for her encouragement to take a chance that I was incredibly
unsure of. Without her encouragement and faith, I would not be here as she had faith in me
when I had no faith in me. I want to thank my brothers, Danny, Brian, and Mark, for setting the
bar ridiculously high on life accomplishments. I appreciate what everyone has done for me and
hopefully I’ve achieved something you can be proud of as I am of all of you. You’ve all helped
me celebrate every success and helped me over every hurdle. My family and friends give my life
purpose and everything I do is to be an asset in each of your lives. Without all of you, this never
would have been possible. From the bottom of my heart, thank you.
iv
ACKNOWLEDGEMENT
During the course of this dissertation, there have been many people who have aided me in
the process. I would like to thank all of the participants who took time out of their schedules to
participate in my data collections. I would like to thank all my lab mates who have been
involved in this process, either through letting me bounce ideas off of them, clarify things with
them, or just generally vent my frustrations periodically, including Tyler Standifird, Hunter
Bennett, Chen Wen, Guangping Shen, Derek Yocum, and especially Lauren Schroeder for all her
assistance in the data collections and processing. Given the amount of time this process takes,
these individuals have made the long hours in the lab much more enjoyable.
I would also like to thank Dr. Harold Cates and Jane Smith of the Tennessee Orthopaedic
Clinic. Without their support and assistance for the project and the subject recruitment, this
project would not have been possible. Additionally, I would like to thank the International
Society of Biomechanics for providing the financial support for this study.
My committee members served an invaluable purpose on this dissertation. I would like
to thank Dr. Josh Weinhandl, Dr. Jeff Reinbolt, and Dr. Rebecca Zakrajsek. Each one of them
has taught me various aspects of the research process and contributed to my general knowledge
of our field, without which I would not be where I am today.
Finally, I would like to thank my advisor, Dr. Songning Zhang. For the last three years,
he has been a great teacher and support system for my professional growth. I have learned a
great deal about the research process and about the mentoring process. I strive to apply this
knowledge to my own future so that I may be a source of pride for him. Your tireless effort on
ensuring my success means the world to me. Thank you for taking a chance on me and for
everything that you have taught me.
v
ABSTRACT
Total knee replacement (TKR) patients have shown alterations in lower extremity
biomechanics during level ground walking and stair negotiation, strength levels, and balance
abilities, however, it is unknown how dissatisfied TKR patients compare to satisfied TKR
patients in these activities. Study One examined the lower extremity biomechanics of
dissatisfied and satisfied TKR patients during level ground walking. Study Two investigated
knee biomechanics during stair ascent and descent activities. Study Three compared isokinetic
strength, balance abilities, deep knee flexion abilities, and functional abilities of the dissatisfied
patients to the satisfied patients. Study Four performed a logistic regression as a means of
examining significant variables in models designed to predict patient satisfaction.
Study One found reduced 1st and 2nd peak VGRF, knee flexion ROM, and peak loading-
response knee extension and abduction moments in the dissatisfied patients compared to healthy
controls. First and 2nd peak VGRFs and flexion ROM were reduced in the replaced limb of the
dissatisfied patients compared to their non-replaced limb. Study Two showed reduced 2nd peak
VGRF and loading-response knee extension moments in the replaced limb of the dissatisfied
group compared to their non-replaced limb and to satisfied and healthy groups during stair
ascent. 1st peak VGRF and both loading-response and push-off abduction moments showed
reduced values in replaced limbs compared to non-replaced limbs for all groups. During stair
descent, the dissatisfied group showed reduced loading-response and push-off knee extension
moments in their replaced limb compared to their non-replaced limb and the healthy group. The
loading-response knee extension and abduction moments were also reduced in the dissatisfied
group compared to the satisfied group. Study Three showed reduced peak extension (180°/s) and
flexion (60°/s) torque in dissatisfied patients compared to satisfied patients. No balance
vi
differences were evident, although an increased percentage of dissatisfied patients were unable to
complete the unilateral balance tests. Study Four produced models via the logistic regression
analysis which often included peak VGRFs and knee extension moments. Future research
should examine the effects of attempting to alter the physical differences between patient
satisfaction groups and whether it improves patient satisfaction rates.
vii
TABLE OF CONTENTS
CHAPTER I INTRODUCTION ..................................................................................................... 1
Background ................................................................................................................................. 1
Statement of the Problem ............................................................................................................ 6
Research Hypotheses................................................................................................................... 7
Significance ................................................................................................................................. 8
Delimitations ............................................................................................................................... 8
Limitations ................................................................................................................................ 10
CHAPTER II LITERATURE REVIEW ...................................................................................... 12
Introduction ............................................................................................................................... 12
Comparison of Motion Capture Methods, Markers, and Calculations ..................................... 13
Two-dimensional vs Three-dimensional Motion Capture ..................................................... 13
3D Marker Sets ...................................................................................................................... 17
3D Joint Angle Measurement ................................................................................................ 25
Biomechanics, Strength, and Balance of Total Knee Replacement Patients ............................ 34
TKR Patient Biomechanics of Level Walking and Stair Climbing ....................................... 36
Spatio-Temporal Gait Variables ............................................................................................ 38
Ground Reaction Force .......................................................................................................... 41
Sagittal Plane Kinematics ...................................................................................................... 43
Sagittal Plane Kinetics ........................................................................................................... 46
Frontal Plane Kinematics....................................................................................................... 50
Frontal Plane Joint Kinetics ................................................................................................... 53
TKR Patient Knee Strength ................................................................................................... 57
TKR Patient Balance Abilities .............................................................................................. 62
Creation, Examination, and Application of Survey Tools ........................................................ 66
Psychometric Properties ........................................................................................................ 67
Development of Surveys Used on TKR Populations ............................................................ 72
Application of Surveys to TKR Population ........................................................................... 91
Patient Satisfaction .................................................................................................................... 92
Meaning of Satisfaction ......................................................................................................... 92
Psychological Factors of Satisfaction .................................................................................... 96
viii
Patient Satisfaction for TKR Procedures ............................................................................... 97
Closing Statement ................................................................................................................... 104
CHAPTER III OVERGROUND WALKING BIOMECHANICS OF SATISFIED AND
DISSATISFIED TOTAL KNEE REPLACEMENT PATIENTS .............................................. 106
Abstract ................................................................................................................................... 107
Introduction ............................................................................................................................. 108
Materials and Methods ............................................................................................................ 109
Participants .......................................................................................................................... 109
Instrumentation .................................................................................................................... 110
Experimental Procedures ..................................................................................................... 111
Data Analyses ...................................................................................................................... 111
Statistical Analyses .............................................................................................................. 112
Results ..................................................................................................................................... 113
Discussion ............................................................................................................................... 115
Conclusion ............................................................................................................................... 120
Chapter III Appendix: Tables and Figures .............................................................................. 122
CHAPTER IV STAIR AMBULATION PATTERNS AND ASYMETRICAL LOADING OF
DISSATISFIED TOTAL KNEE REPLACEMENT PATIENTS .............................................. 125
Abstract ................................................................................................................................... 126
Introduction ............................................................................................................................. 127
Materials and Methods ............................................................................................................ 129
Participants .......................................................................................................................... 129
Instrumentation .................................................................................................................... 129
Experimental Procedures ..................................................................................................... 130
Data Analyses ...................................................................................................................... 132
Statistical Analyses .............................................................................................................. 133
Results ..................................................................................................................................... 133
Discussion ............................................................................................................................... 135
Conclusion ............................................................................................................................... 143
Chapter IV Appendix: Tables and Figures .............................................................................. 144
ix
CHAPTER V STRENGTH AND BALANCE DEFICITS AFFECTING PATIENT
SATISFACTION FOR TOTAL KNEE REPLACEMENTS ..................................................... 148
Abstract ................................................................................................................................... 149
Introduction ............................................................................................................................. 150
Materials and Methods ............................................................................................................ 152
Participants .......................................................................................................................... 152
Instrumentation .................................................................................................................... 152
Experimental Procedures ..................................................................................................... 153
Data Analyses ...................................................................................................................... 155
Statistical Analyses .............................................................................................................. 156
Results ..................................................................................................................................... 157
Discussion ............................................................................................................................... 159
Conclusion ............................................................................................................................... 165
Chapter V Appendix: Tables and Figures ............................................................................... 167
CHAPTER VI LOGISTIC REGRESSION ANALYSES REGARDING PATIENT
DISSATISFACTION WITH TOTAL KNEE REPLACEMENT OUTCOMES ....................... 170
Abstract ................................................................................................................................... 171
Introduction ............................................................................................................................. 172
Materials and Methods ............................................................................................................ 174
Participants .......................................................................................................................... 174
Experimental Procedures ..................................................................................................... 174
Data Analyses ...................................................................................................................... 176
Statistical Analyses .............................................................................................................. 177
Results ..................................................................................................................................... 179
Discussion ............................................................................................................................... 180
Conclusion ............................................................................................................................... 185
Chapter VI Appendix: Tables and Figures .............................................................................. 186
CHAPTER VII CONCLUSION ................................................................................................. 188
REFERENCES ........................................................................................................................... 191
x
APPENDICES ............................................................................................................................ 206
Appendix A: Forgotten Joint Score ......................................................................................... 207
Appendix B: Informed Consents ............................................................................................. 208
Appendix C: Physical Activity Readiness Questionnaire ....................................................... 214
Appendix D: Visual Analogue Pain Scale .............................................................................. 215
Appendix E: Demographic Questionnaire .............................................................................. 216
Appendix F: Recruitment Flyers ............................................................................................. 217
Appendix G: Subject Demographics ....................................................................................... 219
Appendix H: Individual Results for Select Variables ............................................................. 222
VITA ........................................................................................................................................... 251
xi
LIST OF TABLES
Table 1. Peak VGRF (BW) and joint moments (Nm/kg), and knee angle and ROM (°): .......... 121
Table 2. Ankle and Hip ROM (°) and Joint Moments (Nm/kg), ................................................ 122
Table 3. WOMAC Scores Pain for individual tests (0-10 VAS) ................................................ 123
Table 4. GRF and Knee Kinematics/Kinetics during Ascent ..................................................... 143
Table 5. GRF and Knee Kinematics/Kinetics during Descent ................................................... 144
Table 6. Symmetry Index for knee kinematics and kinetics during stair ascent ......................... 145
Table 7. Descriptive statistics, functional tests, and survey data. ............................................... 165
Table 8. WOMAC Scores (100mm VAS) and Passive Knee ROM (°). .................................... 165
Table 9. Peak Isokinetic Knee Extension and Flexion Torque (Nm) and Extension and Flexion
Loading Rate (LR; Nm/s). ....................................................................................................... 166
Table 10. Unilateral overall stability index (OSI), mediolateral stability index (MLSI) and
anteroposterior stability index (APSI) and bilateral OSI, MLSI, and APSI stability indices
(one-way ANOVA). ................................................................................................................ 166
Table 11. Deep knee flexion kinematics and kinetics................................................................ 167
Table 12. Pain for individual tests (0-10 Likert). ........................................................................ 167
Table 13. Logistic regression models for TKR patient satisfaction with respect to survey data,
strength, and 3D kinematics and kinetics for over ground walking, stair ascent, and stair
descent. .................................................................................................................................... 185
Table 14. Slope (B), standard error (SE), Odds ratio (OR), and 95% confidence intervals (95%
CI) for predictive variables for top model of strength, functional test, and survey data. ........ 185
Table 15. Slope (B), standard error (SE), Odds ratio (OR), and 95% confidence intervals (95%
CI) for predictive variables for top model of walking data ..................................................... 186
Table 16. Slope (B), standard error (SE), Odds ratio (OR), and 95% confidence intervals (95%
CI) for predictive variables for top model of stair ascent data. ............................................... 186
Table 17. Slope (B), standard error (SE), Odds ratio (OR), and 95% confidence intervals (95%
CI) for predictive variables for top model of stair descent data. ............................................. 186
Table 18. Dissatisfied TKR patient characteristics ..................................................................... 230
Table 19. Satisfied TKR patient characteristics .......................................................................... 231
Table 20. Healthy control participant characteristics ................................................................. 232
Table 21. WOMAC subscales and total scores and passive knee ROM (°) for Dissatisfied TKR
patients. ................................................................................................................................... 233
Table 22. WOMAC subscales and total scores and passive knee ROM (°) for Satisfied TKR
patients. ................................................................................................................................... 234
xii
Table 23. WOMAC subscales and total scores and passive knee ROM (°) for healthy controls.
................................................................................................................................................. 235
Table 24. Functional testing times and over ground walking, stair ascent, and stair descent
velocity for controls and TKR patients. .................................................................................. 236
Table 25. 1st and 2nd peak VGRF for Dissatisfied TKR patients. ............................................... 237
Table 26. 1st and 2nd peak VGRF for Satisfied TKR patients. .................................................... 238
Table 27. 1st and 2nd peak VGRF for healthy controls. ............................................................... 239
Table 28. Knee Flexion ROM (°) and loading-response knee extension moment (Nm/kg) for
Dissatisfied TKR patients. ....................................................................................................... 240
Table 29. Knee Flexion ROM (°) and loading-response knee extension moment (Nm/kg) for
Satisfied TKR patients. ........................................................................................................... 241
Table 30. Knee Flexion ROM (°) and loading-response knee extension moment (Nm/kg) for
healthy controls. ...................................................................................................................... 242
Table 31. Push-off knee extension moment (Nm/kg) and knee adduction ROM (°) for
Dissatisfied TKR patients. ....................................................................................................... 243
Table 32. Push-off knee extension moment (Nm/kg) and knee adduction ROM (°) for Satisfied
TKR patients ........................................................................................................................... 244
Table 33. Push-off knee extension moment (Nm/kg) and knee adduction ROM (°) for healthy
controls. ................................................................................................................................... 245
Table 34. Loading-response and push-off knee abduction moments (Nm/kg) for Dissatisfied
TKR patients. .......................................................................................................................... 246
Table 35. Loading-response and push-off knee abduction moments (Nm/kg) for Satisfied TKR
patients. ................................................................................................................................... 247
Table 36. Loading-response and push-off knee abduction moments (Nm/kg) for healthy controls.
................................................................................................................................................. 248
Table 37. Ankle dorsiflexion ROM (°), loading-response dorsiflexion moment (Nm/kg), and
push-off plantarflexion moment during walking for Dissatisfied TKR patients..................... 249
Table 38. Ankle dorsiflexion ROM (°), loading-response dorsiflexion moment (Nm/kg), and
push-off plantarflexion moment during walking for Satisfied TKR patients. ........................ 250
Table 39. Ankle dorsiflexion ROM (°), loading-response dorsiflexion moment (Nm/kg), and
push-off plantarflexion moment during walking for healthy controls. ................................... 251
Table 40. Hip extension ROM (°), loading-response extension moment (Nm/kg), and push-off
flexion moment (Nm/kg) during walking for Dissatisfied TKR patients ............................... 252
Table 41. Hip extension ROM (°), loading-response extension moment (Nm/kg), and push-off
flexion moment (Nm/kg) during walking for Satisfied TKR patients .................................... 253
Table 42. Hip extension ROM (°), loading-response extension moment (Nm/kg), and push-off
flexion moment (Nm/kg) during walking for healthy controls ............................................... 254
Table 43. Peak knee extension and flexion torque (Nm) for Dissatisfied TKR patients. ........... 255
xiii
Table 44. Peak knee extension and flexion torque (Nm) for Satisfied TKR patients ................. 256
Table 45. Peak knee extension and flexion torque (Nm) for healthy controls ........................... 257
Table 46. Unilateral overall stability index (OSI), mediolateral stability index (MLSI) and
anteroposterior stability index (APSI) for Dissatisfied TKR patients ..................................... 258
Table 47. Unilateral overall stability index (OSI), mediolateral stability index (MLSI) and
anteroposterior stability index (APSI) for Satisfied TKR patients.......................................... 259
Table 48. Unilateral overall stability index (OSI), mediolateral stability index (MLSI) and
anteroposterior stability index (APSI) for healthy controls .................................................... 260
Table 49. Bilateral static and dynamic overall stability index (OSI), mediolateral stability index
(MLSI) and anteroposterior stability index (APSI) for controls and TKR patients ................ 261
1
CHAPTER I
INTRODUCTION
Background
Osteoarthritis (OA) is a joint disease characterized by degeneration of joint cartilage.
Knee OA is one of the most common forms of OA and it is projected to afflict approximately
25% of the population by the year 2030 (1). As the disease progresses and the cartilage wears
away, the knee joint space diminishes, eventually leading to bone on bone contact and
unbearable pain. A common treatment option for this end-stage OA is a total knee replacement
(TKR). As of 2011, there were over 700,000 TKR operations being performed in the United
States each year (2), and an estimated projection of 3.5 million per year by 2030 (3). The
number of procedures continues to increase in individuals under the age of 60 (4), which means
individuals receiving the replacement will have to live longer with the replaced limb.
Restoration of function of the replaced knee is critical given the significant time, pain, and
monetary investment associated with the operation.
The primary goals of TKR operations involve pain relief, improvement in knee joint
range of motion (ROM), improved joint alignment (if necessary, as malalignment can contribute
to degeneration of cartilage), a restored ability to perform activities of daily living (ADL), and a
return to more advanced physical activities for some patients (5). Many of the subjective survey
tools utilized with the TKR population examine patients’ desires and abilities to return to
performing more advanced activities, suggesting that there is an expectation that the surgery will
help the patient to return to these activities. The desired outcomes are not always present though
as some patients continue to have difficulty with function and lingering pain in the replaced knee
joint.
2
Overall, most TKR operations are considered successful as there are often reductions in
pain levels and improvements in ROM in the replaced knee (6-9). Patient satisfaction rates for
the procedure have been reported between 81-97% (10, 11). This leaves a significant portion of
the TKR population as dissatisfied with the outcomes of the replaced knee. Post-operative pain
(12) and functional limitations (13) are commonly reported by dissatisfied patients. This often
results in decreased performance in common clinical tests (such as timed up and go or a sit to
stand test) when compared to healthy controls (14, 15). These tests are often seen as a defining
point for “success” of the operation as they are deemed to determine the restoration of function
for the replaced joint. However, these test results do not sufficiently explain why the TKR
patients are dissatisfied with the TKR outcomes, thereby suggesting that additional research into
the mechanisms of dissatisfaction is needed.
Current research on the dissatisfied population has focused on survey data and minimal
physical testing (16-18). Lab-based biomechanical studies focusing on level walking, stair
negotiation, strength, and balance may provide insight into the movement profiles of the
dissatisfied TKR population. For example, one study has shown no difference in functional test
scores while showing reduced loading response knee extension moments and increased push-off
knee abduction moments in the replaced limb of TKR patients compared to the non-replaced
limbs and healthy controls during stair ascent (19). Despite similarities in functional tests, the
biomechanical analysis was able to show differences in the movement patterns, which is valuable
in movement recovery.
There have been several studies investigating the TKR population as a whole compared
to healthy controls. These studies have examined overground walking, with special focus on gait
velocity (20, 21), sagittal plane knee ROM (20, 22), and frontal plane knee moments (20, 21).
3
Knee replacement surgeries should help to restore normal mechanics during level walking as
well as more demanding and complex activities such as stair negotiation. Different planar
kinematics and kinetics have shown mixed results for TKR patients during level walking. TKR
patients have been shown to have reduced maximum knee flexion compared to healthy controls,
by an average of 6° (15, 22-24), yet some studies have reported similar maximum flexion angles
(25, 26). Internal extension moments have produced mixed results as well, with TKR patients
being reported as producing lower moments compared to healthy controls during a walking task
(26) while yet other studies have shown no difference between the two groups (20, 25). Peak
knee adduction angles for TKR patients have been shown to return to levels similar to healthy
controls, which has been linked to the correction of the frontal plane alignment during the
surgery (27) Despite the alignment correction, mixed results have been shown with respect to
frontal plane moments. There have been no differences between replaced limbs and healthy
controls for the peak internal knee abduction moment (KAM) (26), reductions in first and second
peak KAM values (20), and increased KAM values due to a lack of change from pre-operative
levels which have been shown as increased compared to healthy controls (27, 28). The
confounding results impair the clarity of the situation but more results suggest KAM values
being closer to those of healthy controls rather than increased. It is currently unknown how
dissatisfied patients differ satisfied TKR patients and healthy controls with respect to overground
walking kinematics and kinetics.
Stair climbing, as an example of a more demanding activity on the muscles and joints
(29, 30), is a common ADL for older and younger adults (31). It has been reported as one of the
top five most difficult tasks for individuals over the age of 60 (32). As patients reach end-stage
knee OA, many frequently report difficulty with climbing stairs (33). Stair climbing is also
4
measured by most survey tools [including the Western Ontario and McMaster University survey
(WOMAC) and the Forgotten Joint Score (FJS]) used on the TKR population (5, 34, 35), thereby
suggesting its importance in everyday life for most people. During stair ascent and descent, knee
flexion ROM has been frequently shown to be reduced compared to healthy controls (22, 36, 37),
although one study did report no differences between the two groups (19). TKR patients have
shown reduced knee extension moments during stair ascent, which is often coupled with a
reduction in velocity (24). Gait speed, for example, has been shown to increase post-operatively
for TKR patients compared to pre-operative levels, however, gait speed does not always reach
levels equal to those of healthy controls at one year post-operation (15, 38). Other studies have
shown no differences between the two groups though, with TKR patients maintaining their
velocity from 12-18 months as far as 46 months after surgery (22, 26). As during walking trials,
knee adduction angles have shown no differences for TKR patients compared to healthy controls
but an increased 2nd peak KAM was present (19). Other studies have shown mixed results with
some showing increased KAM values (39) and some showing values equal to or reduced
compared to healthy controls (40-42). For the dissatisfied population, deficits in their movement
profiles may be further enhanced during more difficult activities, suggesting a need for
examination of these activities.
A return of strength levels following an operation is crucial for a return to normal
function. Quadriceps strength is the most commonly measured strength variable for TKR
patients however the hamstrings also have significant function at the knee, warranting their
examination as well. Significant reductions in quadriceps and hamstring strength are evident
early in the rehabilitation process, with upwards of 60% deficits compared to pre-operative levels
(43, 44). By six months post-operation, both muscle groups show significant increases in
5
strength (45). However, the strength levels of TKR patients do not always meet levels of healthy
controls, with deficits still present 12 months after surgery in the replaced limb and the non-
replaced limb not being different from healthy controls (46). There is a lack of research on
strength with respect to patient satisfaction. Strength deficits post-operatively may be more
pronounced in dissatisfied patients, which may impair functional ability.
Balance is an additional measure of success for TKR operations and return to normal
daily activities as falls can be detrimental to the TKR (47). TKR patients have been shown to
have decreased stability after surgery compared to healthy controls (48). Improvements in
balance have been associated with improvements in functional tests, such as stair climb, 30
second chair rise, timed up and go, and improvements in gait speed (49). Strength and balance
have been often measured together as a means of using strength to explain balance abilities.
Increased knee extensor strength coupled with an increased gait speed has led to increased
anterior-posterior (AP) balance (measured through the range of the AP trajectory for the center
of pressure (COP)), but increased knee extensor strength with a reduced gait speed led to a
reduced AP balance (50). However, it was shown that peak torque did not predict single leg
static balance performance (51). Conversely, it has found that a failure to maintain single limb
balance is explained by older age, higher body mass index, and reduced quadriceps strength (52).
Timing may play a role in the analysis of biomechanical variables as differences have
been shown to diminish a year after surgery (53). Other differences have been shown at an
average of 46 months post-operation (22), suggesting that while there may be improved gait
biomechanics through the first year, there may be a regression of improvement over time. The
sub-optimal physical outcomes after surgery may contribute to patient dissatisfaction. An
understanding of a comprehensive physical profile including gait biomechanics, strength,
6
balance, physical functions and patient perceptions of the dissatisfied patients as it compares to
the satisfied patients as well as healthy counterparts is essential for potentially improving the
satisfaction rates of patients.
Statement of the Problem
To our knowledge, no studies have examined the 3D biomechanical profiles, strength
levels, or balance abilities of dissatisfied TKR patients. Simple motion tasks such as level
walking and more demanding tasks such as stair climbing are everyday activities for the TKR
population yet no research has examined these activities specifically for the dissatisfied
population. Therefore the purposes of proposed studies are listed below:
Study One: The purpose was to compare the lower extremity movement profiles of
overground walking for dissatisfied TKR patients to satisfied TKR patients and healthy older-
adult controls.
Study Two: The purpose was to compare the lower extremity movement profiles of stair
ascent and descent for dissatisfied TKR patients to satisfied TKR patients and healthy older-adult
controls.
Study Three: The purpose was to examine the knee flexor and extensor concentric
strength levels of both the replaced and non-replaced limbs for dissatisfied TKR patients
compared to satisfied TKR patients and healthy older-adult controls. An additional purpose was
to compare bilateral and unilateral balance abilities for dissatisfied TKR patients to satisfied
TKR patients and healthy older-adult controls.
Study Four: The purpose was to examine associations between TKR dissatisfaction and
satisfaction and the gait biomechanics, strength, balance, functional capacities, and survey data
7
(measuring joint awareness, pain, stiffness, and functional ability) using a logistic regression
analyses.
Research Hypotheses
Study One
1. It was hypothesized that dissatisfied TKR patients would exhibit increased frontal
plane knee joint moments and reduced sagittal plane knee joint moments in their replaced limb
compared to their non-replaced limb in level walking.
2. It was hypothesized that dissatisfied TKR patients would exhibit increased frontal
plane knee joint moments and reduced sagittal plane knee joint moments compared to satisfied
TKR patients and healthy controls in level walking.
Study Two
1. It was hypothesized that dissatisfied TKR patients would exhibit increased frontal
plane knee moments and decreased sagittal plane knee moments in their replaced limb compared
to their non-replaced limb, but similar in the non-replaced limb to the satisfied TKR patients and
healthy controls during stair ascent and descent.
Study Three
1. It was hypothesized that dissatisfied TKR patients would show a knee extensor
strength deficit in their replaced limb compared to their non-replaced limb and a deficit in both
limbs compared to satisfied TKR patients and healthy controls during concentric isokinetic
strength testing.
2. It was hypothesized that dissatisfied TKR patients would have a reduced ability to
balance on their replaced limb compared to their non-replaced limb, satisfied TKR patients, and
healthy controls.
8
Study Four
1. It was hypothesized that reduced knee extensor moments, reduced quadriceps
strength, increased balance stability indices, and increased pain will contribute to increased
dissatisfaction with TKRs.
Significance
Current research regarding the dissatisfied TKR patient population lacks comprehensive
examination of the objective data with respect to gait biomechanics and strength and balance
abilities. This presents a unique opportunity to potentially identify physical differences between
satisfied and dissatisfied patients, which could potentially help to identify treatment objectives
for the dissatisfied population and thereby improve satisfaction rates. Deficiencies in gait,
strength, and balance are often modifiable and do not require surgical intervention. If there is a
potential non-surgical intervention available to help improve patient satisfaction, then it should
be explored as the commitment to a TKR procedure is a significant one and therefore should
have the optimal outcomes whenever possible. Additionally, the results from this study may
provide indirect feedback for the improvement of future TKR implant designs.
Delimitations
The exclusion criteria included for dissatisfied and satisfied TKR patients for all studies:
Diagnosed osteoarthritis at the ankle, knee (contralateral knee of the TKR side), or hip
joint as reported by the patient.
Any additional lower extremity joint replacement.
Any lower extremity joint arthroscopic surgery or intra-articular injection within past 3
months.
9
Systemic inflammatory arthritis (rheumatoid arthritis, psoriatic arthritis) as reported by
the patient.
BMI greater than 38.
Neurologic disease (e.g. Parkinson’s disease, stroke patients) as reported by the patient.
Any additional major lower extremity injuries/surgeries aside from knee replacement.
Inability to walk without a walking aid.
Any visual conditions affecting gait or balance.
Women who are pregnant or nursing.
Any cardiovascular disease or primary risk factor which precludes participation in
aerobic exercise as indicated by the Physical Activity Readiness Survey. If any
participant marks “yes” on the survey they will be required to obtain written consent from
their doctor indicating they are healthy enough to participate in the study.
An answer of “neutral” on satisfaction question in pre-screening interview.
The inclusion criteria included for dissatisfied and satisfied TKR patients for all studies:
Men and women between the ages of 50 and 75.
Total knee replacement in one knee.
At least 12-months from TKR.
No more than 5-years from TKR
The exclusion criteria included for healthy adults for all studies:
Knee pain for at least 6 months during daily activities.
Diagnosed with any type of lower extremity joint osteoarthritis (self-reported).
Any lower extremity joint replacement.
10
Any lower extremity joint arthroscopic surgery or intra-articular injection within past 3
months.
Systemic inflammatory arthritis (rheumatoid arthritis, psoriatic arthritis) (self-reported).
BMI greater than 38.
Inability to ascend/descend stairs without the use of a handrail.
Inability to walk without a walking aid.
Neurologic disease (e.g. Parkinson's Disease, stroke patients) (self-reported).
Any major lower extremity injuries/surgeries.
Any visual conditions affecting gait or balance.
Women who are pregnant or nursing.
Any cardiovascular disease or primary risk factor which precludes participation in
aerobic exercise as indicated by the Physical Activity Readiness Survey. If any
participant marks “yes” on the survey they will be required to obtain written consent from
their doctor indicating they are healthy enough to participate in the study.
The inclusion criteria included for healthy adults for all studies:
Men and women between the ages of 50 and 75.
Limitations
All studies were conducted in a laboratory setting.
Skin marker placement on obese patients may not reflect accurate bony landmark
locations given excess tissue
Reflective markers used to track the feet during motion trials were placed on the shoes,
and therefore motion of the foot within the shoe may not have been accurately captured.
11
The nature of the staircase set up required placement prior to subjects arriving for testing,
therefore level walking tests always occurred after stair ascent and descent.
There was no training period for the strength and balance tests although practice trials
were provided.
An isokinetic speed of 180°/s may be too fast of a velocity for some TKR patients to
contract at, thereby altering torque values.
The foot placement on the balance system is set by the Biodex and does not allow
participants to balance as they naturally would with whatever foot placement they desire.
12
CHAPTER II
LITERATURE REVIEW
Introduction
The purpose of this study was to examine the gait biomechanics, strength, and balance
profiles of dissatisfied total knee replacement (TKR) patients and how they compare to satisfied
TKR patients and healthy adults. Information on TKR patient groups has previously been
reported with respect to gait biomechanics, strength, and balance, however previous research has
failed to disseminate this information based upon patient satisfaction with the TKR procedure.
The dominant information relating to the satisfaction grouping has been based upon widely
distributed survey data, such as the Western Ontario and McMasters University (WOMAC)
survey, which quantifies the pain, stiffness, and function levels of TKR patients. All of this
information is self-reported by the patient as is consistent with the patient satisfaction reporting.
To date, no study has examined the non-subjective outcomes such as gait biomechanics, strength,
and balance profiles to see how these factors relate to the subjective patient assessment of
satisfaction.
The primary goals of this chapter were to 1) examine the available motion capture
methodologies for biomechanical profiling, 2) summarize the current findings related to TKR
patients for gait biomechanics profiles, strength, and balance, 3) examine the creation and
application of survey tools which are commonly utilized on the TKR population, and 4) examine
the current construct of patient satisfaction and the literature on dissatisfied TKR patients for the
purpose of identifying the gap in the available literature.
13
Comparison of Motion Capture Methods, Markers, and Calculations
Motion capture has become a regularly used tool for the analysis of human movement.
Everything from healthy gait to joint replacements to neurological disorders have been studied
for their movement deficiencies using various motion capture systems (54, 55). Initially motion
capture study was conducted in 2D using a series of cameras to take pictures. Since then,
methods and software have evolved to provide a more thorough, albeit complex, understanding
of how the body moves through the use of multiple cameras, recognition software, tracking
systems, and various imaging tools. This has provided for a more intimate and in-depth
understanding of physical function as it relates to movement patterns. The evolution of motion
analysis has taken several years and examined several different pieces of equipment in the
pursuit of the most accurate assessment of human motion. The goal has been to develop a
system that is easily applicable, minimally invasive, and provides an in-depth, accurate analysis
of movement with no errors, a concept which likely still eludes the research community.
However, many strides have been achieved in this field as it continues to evolve.
Two-dimensional vs Three-dimensional Motion Capture
There are multiple motion capture methods currently available for use in research.
Three-dimensional (3D) has become the most common in the laboratory setting, however, two-
dimensional (2D) is also currently being used in field and clinical settings when 3D is not as
readily available. 2D refers to a projection of body segments onto a single plane of motion while
3D models the body using simplified 3D objects such as cylinders or superquadratics. There is
also a system between 2D and 3D called 2.5D which provides some depth information but not at
the level of 3D (56). Each system has its own benefits and setbacks.
14
In general, 2D data is much easier to obtain than 3D as it only requires one camera and a
few markers (57). The camera must be placed perpendicular to the plane of motion being
studied, a few markers can be applied to the subject to define end points of the segments, and the
data can be processed to calculate kinematics (and kinetics if a force plate is combined). This
simple data collection can provide useful information with respect to movement in the studied
plane. In addition to the minimal equipment, the research can be conducted in almost any
environment as the equipment does not take up much space and there are not many additional
pieces of equipment to accompany the single camera. This allows for field research much easier
than 3D. There are currently several commercial software applications which also allow for
kinematic measurements using video from cellular phones which requires even less set up and
processing. The 2D data can be further enhanced using silhouette modeling which helps to track
the subject through the movement using a constraint mechanism on the model and a weighting
scheme on the segments of the model to provide more in-depth data (58), although this requires
some more initial work to process the model.
The ease of application in the 2D data is offset by some limitations. First, and most
importantly, most human motion is multi-planar. While some motions are largely occurring in
one plane (walking in the sagittal plane), the reality is that humans move in all three planes
during everything that they do. The assumption of planar motion in 2D data collection is simply
not realistic, especially in the more dynamic motions such as dynamic stability during a slip and
fall (59). The non-sagittal plane motions are often small, but important nonetheless, and 2D
through a single camera can miss these elements. Additionally, the data can be skewed if the
camera is not precisely set up, movement occurs out-of-plane (which induces an error in the
data), or if the subject moves off the perpendicular line to the field of view (58). If the sagittal
15
plane is the plane of interest, failure to set up a perpendicular field of view to the plane can result
in inaccurate data as the perception of the subject changes, which alters the tracking of the
subject in the data. The main limitation, as far as common occurrence, is the occlusion of
markers (58). When utilizing only one camera and minimal numbers of markers, it can be easy
for the markers to become blocked from the camera’s field of view, such as an arm swinging in
front of a hip marker during walking. This will affect the kinematic data as there is no additional
camera to compensate for the temporary loss of vision of the marker, resulting in two possible
options. First, the marker could be gap-filled based on the assumed trajectory the next time it
comes into view digitally. This may not be accurate, as realistically there is an element of
guessing done by the software which fills the missing time points. Second, the researchers
tracking the data points do their own guessing of approximately where the marker should be.
This is less-likely to be accurate, but may be the only option with some of the commercial
applications. When a rough idea of kinematics is all the information desired by the researcher,
this may not be a severe limitation, but if there are clinical implications to the changes in body
position, this becomes a more significant problem.
The 2D collection can be improved upon through the use of 2D fluoroscopy. These 2D
images can be transformed into 3D models of the bones for kinematic measurements (60).
Through the use of a 3x3 rotation matrix followed by a 3x1 translational vector matrix,
fluoroscopy has resulted in a root mean square error (RMS) of 0.7 mm (61), making the
measurements very accurate. The maximum errors in the fluoroscopy were found to be in-plane
marker translational errors of 1.5 mm, for translations normal to the image plane, 3 mm, and 0.6°
for rotations in all planes (60).
16
3D data collection does improve upon some of the issues with 2D. In order to
successfully obtain 3D data, there must be at least two cameras capturing at least three markers
on a segment in order to take the 2D coordinates and reconstruct the markers in 3D space (62).
The 3D coordinates of the markers are computed from the 2D marker coordinates of each camera
and then matched stereometrically into 3D space (63). This allows for assessment of human
motion in all planes, which is more realistic as research has shown motion exists in all three
planes simultaneously (64). Proper camera placement is still important but unlike 2D where
precision is required, the use of multiple cameras requires less “precise” set up. As long as two
of the cameras in the system capture the marker, it is reconstructable in 3D space. This further
helps to eliminate occlusion of markers as there are more cameras capable of capturing the
marker information from a multitude of angles.
There are limitations however as the equipment can be costly. Unlike 2D where one
camera and only a few markers are needed, 3D requires multiple cameras, often as many as 12
are utilized (65). These cameras often require additional expensive software packages for the
tracking and processing of the data. Costs aside, the bulk of the equipment makes it difficult to
take this type of system out of a laboratory setting (66), rendering much of the field-based data
collections impossible due to the complexity of the transportation and set up. While the
laboratory is an acceptable environment for some research, for athletic tasks, environment plays
a substantial role and the laboratory setting negates that, taking away an element of reality and
creating a limitation on the data results. An additional limitation for both 2D and 3D is the
clothes worn by the subject as they need to be minimal or tight fitting so as to minimize marker
motion related to the underlying bones they represent (66). Loose fitting clothing can cause the
markers to swing off of the represented bones, causing inaccurate kinematic calculations. There
17
are tracker programs which can be combined with as little as two cameras in creating 3D data. A
model is created based on the subject and attempts to follow it through the motion which can
help to eliminate occlusion errors, but most tracker programs have been shown to perform better
with the use of more than two cameras (56). Beyond this, recent research has shown that
construction of 3D images are possible through the use of one camera and a retro-grade reflector
system. This has shown high correlations with movement in all three planes compared to
traditional 3D motion capture (67).
When directly comparing the 2D and 3D data collections, the two methods have shown
similar joint angle and moment profiles. A 2D versus 3D analysis of the hindlimbs of horses
revealed no significant differences in kinematics during a handled trot motion (68). However,
the magnitudes of the joint moment profiles were significantly different (despite the similarity of
the motion profile pattern). The 3D data exhibited increased dorsiflexion moments, peak knee
extensor and hip flexor moments in the second half of stance, decreased knee flexor moments
during midstance, and no difference for hip extensor moments (62). The differences could be
eliminated for the ankle and hip by using the 3D joint center positions for the 2D calculations,
suggesting some sort of offset in the 2D calculations to account for the differences, however
even with the change in calculations, differences still existed in the knee joint (62). The
calculation differences have yet to be substantiated as far as an acceptable standard is concerned.
This means that direct comparisons between 2D and 3D data should be made with caution as the
magnitude differences may have significant implications regarding the studied populations.
3D Marker Sets
As motion capture technology has evolved, so too has the marker sets utilized in the data
collections. 2D markers were often placed at segment end points and/or joint centers from the
18
perspective of a perpendicular viewpoint to the plane of motion. 3D data has surpassed this in
terms of total markers utilized with a minimum of three markers needed to create a segment’s
position in space (62). This has led to the question of whether the markers being used truly
represent the underlying bone motion which they are supposed to mimic. There are two
dominant methods of skin-based markers being utilized: the Helen Hayes marker set (69) which
is a minimalist approach designed to use fewer markers and a Cleveland Clinic cluster-based
approach (70) which uses more markers in an attempt to model the underlying bone motions.
Skin marker-based motion capture is the most commonly used method in human motion
research, however prior to the in-depth investigation into the different skin-based marker sets, it
should be mentioned that neither of these represents the “gold standard” for marker-based
motion capture. Bone pins have been identified by several researchers as the “gold standard” for
marker-based motion capture (71-78). Bone pins are the surgical insertion of a pin into the bone,
with the end of the pin containing the same markers used in skin-based marker motion capture.
The same motion capture system is used but the bone pins are believed to represent the actual
bone motion since they are directly attached and do not have to contend with movement of the
skin and muscles which separate the skin markers from the bones. Bone pins are subject to error
though as the bending of pins or impingement can occur as the muscles and tissue still has to act
around the pin. This can be alleviated by using shorter threaded pins and fully inserting the
threaded part into the bone as the smooth part of the pin is more resistant to bending as are larger
diameter pins. Additionally, specific placement of the pins between muscles/tendons where the
least amount of movement occurs is beneficial. Dynamic motion on the operating table may help
to better guide the pins into the correct position (76). This method is infrequently used due to the
invasive, surgical nature of the pin attachments, which would dissuade many potential
19
participants from being involved in the research, not to mention the added expense of the
surgical procedure with respect to time, equipment, anesthesia, etc.
Given the complexity of the bone pins, the Helen Hayes and Cleveland Clinic skin
marker sets have dominated the research protocols. In beginning with the minimalist approach
(Helen Hayes), this marker system was initially developed for low resolution imaging systems
with a goal of having as few markers as possible spaced as far apart as possible (79). Markers
were placed bilaterally on the anterior superior iliac spine (ASIS), the top of the sacrum, greater
trochanter, lateral aspect of knee flexion axis, lateral malleolus, and between second and third
metatarsal heads (69). This was done in an attempt to add as little alteration to the natural
subject as possible, which is inherently beneficial to simulating realistic research as people do
not move with markers on their body in their day to day lives. This marker set constrains the
joints with three rotational degrees of freedom (DOF), it creates a thigh segment that relies on a
hip joint center estimation from pelvic markers, creates a thigh that shares a knee joint center
marker with the shank, and has a foot defined by an ankle joint center created from shank
markers (64). Shared markers inherently provide a limitation as joints are not a connection point
between two bones, per se. The knee, for example, is not the connection point between the
femur and tibia. There is a space that joins the two bones together to create the joint, but they are
often modeled as the distal point of the femur being the same as the proximal point of the tibia,
which is part of the rigid body model assumption. The axis of rotation for the knee sits in the
femoral epicondyles, but user placement errors of markers can be a limitation.
One issue with the minimalist approach is that it is modified quite frequently with respect
to placement of the markers. In theory, if the same tests are performed by different researchers
on the same subject, the results should be the same. This is not always the case though as inter-
20
researcher reliability is not always high. In 2004, Schwartz et al performed an analysis using
four different therapists in three different sessions using a modified Helen Hayes marker set.
They all used the same marker set, but results showed differences in the examined kinematics.
There were higher errors for pelvic tilt than the other pelvic rotations. The frontal plane hip
motions were the most reliable with the transverse plane hip motions providing the largest errors.
The frontal plane knee had the smallest inter-trial errors, but highest inter-therapist to inter-trial
ratio, suggesting experimental errors. The foot progression angle had large inter-trial and inter-
therapist errors, which was attributed to improper foot marker placement (65). This could prove
problematic for a pathology that needs to examine foot rotation or desires to correct the
pathology through a movement intervention.
It has been suggested that more markers are needed with this minimalist approach in
order to improve the model. By expanding the markers, the loss of markers can be minimized,
which is especially important during faster motions where marker tracking difficulty increases
(56). The three DOF is also seen as a problem. A 6 DOF model has been shown to have better
construct validity combined with fewer theoretical assumptions embedded in the model,
including a lack of joint constraint and independence between segments (64). When comparing
the simultaneous measurement of the Helen Hayes markers with a 6 DOF model, altered
kinematics were observed. These differences included magnitude differences (frontal/transverse
pelvis, frontal hip, knee external rotation) and pattern differences, often with opposing trends
(transverse hip, frontal knee) and changes in time occurrence of peaks during stance and swing
phases (64). There was increased inconsistency in the frontal plane knee kinematics, suggesting
cross talk in the Helen Hayes set with knee flexion, characterized by anatomical identification
difficulties (64). The anatomical landmark identification can also be problematic with different
21
types of patients, as increased tissue presence can make identification of the landmarks tough.
Given the inconclusive results of the comparisons of the Helen Hayes set and the 6 DOF model,
it was suggested that the joint constraints are not the greatest limitation of the Helen Hayes set.
However, they are still a significant limitation, meaning that since the results of 6 DOF model
are not worse, this would be a better starting point for research than the Helen Hayes set, despite
the soft tissue artifact (STA) problems for the 6 DOF model (64).
In opposition to the Helen Hayes set and congruent with a 6 DOF model, the Cleveland
Clinic marker set is characterized by a triad of markers attached to a rigid shell and positioned
over the lateral aspect of the body segment, allowing for the creation of 6 DOFs in the rigid body
model (70). This shell is for tracking the movement while proximal and distal landmarks are
chosen to estimate the segment end points for building the rigid body model (80). There are
different possible placements for the marker clusters, each of which have been attributed to their
own problems. Mid-lateral aspect of the segment tends to be the most common placement, but
the distal-lateral segment has been argued as the most accurate placement of the cluster as far as
reproducing accurate movements, for both the cluster and the Helen Hayes wand marker (80).
The various cluster arrangements possible have shown good agreement within the sagittal plane
(81) but movements outside of the sagittal plane are more susceptible to errors. Similar
biomechanical patterns were observed and it was hypothesized that this may be linked to a bias
in the axis of rotation and related cross-talk between markers rather than poor positioning of the
markers (81). They reproduce kinematics well but do not appear to represent the bones well with
the given translational and rotational errors (73). This methodology has also been shown to be
susceptible to further errors based on their attachment methodology. There is the potential to
wrap over the markers, under the markers, or several other methods. An underwrap
22
methodology was shown to be the best method for securing the cluster on the shank, as it was for
the wand-projected marker on the Helen Hayes set (80).
There have been multiple comparisons of different marker sets in an attempt to examine
errors related to the incorrect motion of the skin-markers to the bones they are targeted to
represent. Most 3D data collections with skin-markers treats the body as a rigid body model
where the movement of the markers represents the underlying bones, however STA has been
shown to be a problem. There is a substantial amount of tissue surrounding the bones and as the
body moves, muscles contract and lengthen and can change the positioning of the markers (71),
providing a false reading of movement related to the bone. This is largely subject dependent
though as there are many different body shapes and sizes which can result in different levels of
change (73). In an examination of the lower limbs, STA has been shown to be higher on the
thigh than on the shank during open chain knee flexion, hip axial rotation, level walking, step up,
land and cut, and stair climbing (60, 82-85). The level of STA is based on the placement of the
markers as well as the motion being performed. During a step up activity, the largest mean error
related to STA of the thigh was 12.6 mm in the proximodistal direction and during a walking
activity, it was 19.1 mm in an anteroposterior direction (60). On the shank, it was during open
chain knee flexion with an error of 8.6 mm in the mediolateral direction (60) and during walking
it was 9.3 mm in the anteroposterior direction (82). Errors have been shown to be as high as 40
mm (86). Many of the factors in motion analysis interact, for example, where errors of lower
than 3 mm have been shown to have no effect on joint angles over a 30° flexion range of motion
(69). Large knee flexion motions have shown mid-anterior thigh markers moving distally to the
underlying bone (60). This makes sense as the knee flexes, the skin could be pulled in a distal
direction, taking the skin-mounted markers with it in the same direction. In a similar manner, a
23
patellar marker had larger RMS errors than any of the thigh markers in the proximodistal
direction (60), likely due to movement of the quadriceps muscles which has a direct connection
to the patella and is therefore very susceptible to every contraction of those muscles.
These errors can also be phasic, rendering correction factors difficult to implement.
During walking, the thigh markers can experience a proximal and anterior shift after heel strike
(60). This can manifest as temporary inaccuracies, as it only occurs at certain points. Phasic
contractions may contribute to increased STA during gait (77). At 10%, 50%, and 100% of a
gait cycle during running, inaccuracies have been reported (86). Most of the STA errors
affecting kinematics and kinetics are during the first half of the stance phase (85). The
complexity of the movement also impacts it. More dynamic motions increase the error values.
When comparing walking to a cutting motion, translational errors have been shown to increase
from a range of 3.3-13 mm to 5.6-16.1 mm and rotational errors from a range of 2.4-4.4° to 3.3-
13.1° (73). The rotational errors have reached values as high as 19° in the transverse plane
during walking (82). As the flexion values increase, the non-sagittal angle errors increase
meaning that caution should be taken when comparing different data sets which represent
different flexion levels (69).
As a result of these errors, it has been suggested that standard errors be applied for the
interpreting of kinematics (73). The issue with this is the large discrepancy in the errors and how
to create a standardized calculation for this. Multiple calibration poses have been used in an
attempt to alleviate this but have still resulted in translational and rotational errors of 3 mm and
3°, respectively (87). Standardization of marker placement may be an option to help alleviate
some of these issues with differences between data sets. The lateral aspect of the thigh and
shank are believed to cause minimal errors and therefore the best results (86). Rotational
24
deviations are the smallest when placed laterally and distally, with error values less than 3° in the
sagittal and frontal planes, although that number represents 2-3% and 10-25%, respectively, of
the total range of motion (80) and should therefore still be interpreted with caution. STA errors
for the knee joint have been reported between 2.4-8.3°, but often the non-sagittal values exceed
the total range of motion (60). This is on the higher end of the spectrum for absolute values
compared to other studies (84), but was likely due to a larger range of motion examined.
These errors may be reduced through certain data processing techniques. An interval
deformation technique based on a point cluster technique (88) that can utilize the mathematical
concepts to reduce the position and orientation errors of skin markers by 33% and 25%,
respectively, in the rigid body model (71). The errors largely depend on the base of reference
though. For example, if fluoroscopy is the “gold standard”, the RMS errors for the skin markers
can reach upwards of 190% (84). Additionally, the sagittal and frontal moments of the skin
markers are reported as lower than the fluoroscopy (61).
Irrespective of the marker set used, the biggest errors reported are in the non-sagittal
planes. There is good consistency with multiple marker sets in the sagittal plane for the lower
limb joints, but the correlations significantly decline once out of the sagittal plane (81, 84).
Transverse plane deviations tend to occur during early and late stance (80), lending some support
to transverse validity during mid stance. These deviations during the first third of stance may be
related to inertial effects of the segments related to heel strike (77) and during the last third may
be related to increased shank muscle activation in an attempt to propel the body forward (89).
Marker placement is one of the biggest sources of error between labs, with inter-lab differences
as high as 26% (pelvic rotation) and 33% (foot progression angle), although most differences
were still 8% or less (90). Eight percent can still represent a significant error when comparing
25
data which may have clinical implications. The altered skin marker motion leads to altered joint
centers, which can significantly impact the moment arms and joint moments (61). This in
conjunction with altered joint angles can impact the interpretation of the data. It has been
suggested that data errors in excess of 5° may be large enough to mislead clinical interpretation,
between 2° and 5° may be acceptable but requires consideration in interpretation, and smaller
than 2° are too small and do not require independent consideration in result interpretation (91).
However, as noted earlier, this may represent high percentages of total range of motion. It is
believed that a compensation method is necessary and should be both subject and task specific
(84), although a generally accepted method has yet to be presented.
One final compensation is a markerless design built to construct 3D images based on 2D
camera spatial recognition. The advantages of this system being that no initialization is needed,
the subjects can go from the static pose to immediate motion, and there is a direct provision of
joint centers and segment volume information during motion to allow for better kinematic
calculations (92). However, extensive research has not been completed with this system, leaving
discussion regarding its validity and mass application.
3D Joint Angle Measurement
Currently, a frequently used construct for the calculation of 3D joint angles is based on
the rigid body model, which as described earlier is with limitations due to the muscles activating
and moving the markers (71). However, as this is the commonly utilized system, joint angles are
based within this system. There are a multitude of methods for calculating the joint angles based
on the marker data. Much of the current research is based on Euler and Cardan angles which are
rotations about a given axis as defined in either the global or local systems. There are multiple
orientation possibilities within each system, each of which ultimately provides 3 angular
26
numbers which reflect the rotation from the initial static pose (57). Human bodies which are
modeled as rigid bodies have rotations assumed to take place around a fixed point. Euler angles
have been applied in a manner that allows for the description of one segment rotating about
another segment in a 3D space (93). These angular calculations require orthogonal embedded
axes for both the moving segment and the reference segment. In the proximal reference frame,
unit vectors are applied to rotation matrices from derived equations, although assumptions can be
made with respect to the order of the rotational occurrences (69). When the order assumptions
exist, the initial axial definition is important. For example, in a sagittal, frontal, transverse order,
the sagittal axis can be defined by the skin markers. These markers are placed in a manner
where skin movement is minimized around the underlying bone in an attempt to satisfy the rigid
body assumption. Correct positioning of these markers is crucial though as all subsequent
rotations are affected by the placement of these markers and subsequent axis definition (69).
When the initial axis is created based on anatomical landmarks, a line is drawn between the two
markers to create the initial axis. For example, the flexion/extension axis of the ankle can be
created through joining the medial and lateral malleoli markers. From this midpoint (the ankle
joint center), a line is drawn to the knee joint center (which is often the midpoint between the
femoral epicondyles or some other anatomical landmarks around the knee) to create the
longitudinal axis. The third axis is the cross product of the other two axes which creates the
anteroposterior axis (75), although this may not create two axes orthogonal to each other and it
therefore cannot be used for Euler/Cardan angle calculations. This gives an orientation about the
ankle joint center but this orientation can be changed to reflect a different segment, although this
has been shown to alter joint moments in terms of magnitude (sagittal/frontal) and profile
(frontal) (75). Singular value decomposition method also utilizes a position vector and
27
orientation matrix which can be obtained from the transformation matrix that was estimated by
an algorithm connecting the local coordinate frame of three skin markers (tracking markers) and
their global positions (78, 84, 94).
Each coordinate system has its own mathematical constructs which lead to different
results. The reported coordinate system for joint angles is a significant variable to take into
account. Joint coordinate systems (JCS), global coordinate systems (GCS), and local coordinate
systems (LCS) are all utilized on a regular basis within the research community. Each is chosen
for its respective application to the research being conducted, however each needs to be carefully
analyzed. The non-orthogonal JCS was initially proposed by Grood and Suntay (95). In order to
set up the JCS, a Cartesian coordinate system is established for adjoining segments. The axes in
this system are defined either by palpable anatomical landmarks or through X-ray imaging. The
common origin of both axis systems for the segments becomes the point of reference for the
linear translation occurring in the joint. The JCS is then established based on the intersection of
the two Cartesian coordinate systems. Two of the axes are fixed in the body while one is
“floating”. This provides for three rotational and three translational joint motions within the JCS
(96). Kinematic data can then be calculated based on the rotational elements, feeding into the
calculation of net internal joint moments through a standard inverse dynamics approach.
Calculations of external moments using two components for independent effects is also possible:
a moment due to the resultant ground reaction force (GRF) and moment arm and a free moment
of rotation about the vertical axis through the center of pressure (97). Using the JCS facilitates
correspondence between kinematics and kinetics. A projection onto the JCS is suggested as the
only method where joint moments correspond to muscle and ligament loading and may be the
natural choice for a standardized procedure (98).
28
The JCS proposed by Grood and Suntay is the same as a Xyz Cardan rotation sequence,
where “X” represents a flexion/extension rotation axis, “y” represents an ab/adduction axis, and
“z” represents an internal/external rotation axis, which is commonly used in biomechanics
research. However, this is not the only available rotation sequence. Cardan and Euler rotations
allow for different methods of calculating joint angles, based on different rotation patterns.
There are three potential rotations which can be sequenced: X, Y, and Z. These can be placed
into different orders of operation which allow for different angle calculations. The resulting
angles are known as Cardan angles when the rotation sequence has only one of each letter (i.e.
Xyz, Zxy, etc.). When a letter is repeated in the sequence (i.e. Xyx, ZxZ, etc.), it is known as
Euler angles (99). This allows for the creation of three independent angles, however the
magnitudes of these angles change depending on how the rotational sequence is ordered, thus
confining the angles to a “sequence dependency” (100).
An examination of the different rotational sequences has provided some consistent results
and some mixed results. During an assessment of lower extremity dynamic tasks, the different
Cardan sequences provided good agreement for motion in the X axis, except for during a YXZ
sequence (101). This led to a conclusion that the X axis is robust in its usage during the Cardan
sequences. Upon choosing a following rotation from Y and Z axes, results become mixed with
numerical offsets and profiling changes in some instances (101). Some rotation sequences which
have shown higher RMS values may have been due to higher flexion values which can affect the
joint center and axis of rotation thereby causing errors in following computations (101). One
potential problem with the high angle values is the potential for Gimble lock, which occurs as the
second orientation angle approaches 90° (102). In the lower extremity, the flexion/extension
motions are typically the only motions which approach (±10°) that value. Therefore
29
implementing the X axis as the second rotation sequence may not be in the best interest of
accurate calculation as the mathematical singularity becomes an issue and can cause
compounding errors (103). This is further supported by a direct comparison of the different
Cardan sequences which showed that the YXZ and ZXY rotation sequences performed the worst
(101). In a study examining a Euler rotation sequence beginning with X, there was good
agreement with a JCS rotation, but when the Euler sequence began with Z rotation, the
kinematics were significantly altered in comparison to the JCS (104).
The rotational sequence should be based on the movement of interest as well as where the
largest motions occur. In most lower extremity motion, the bulk of the movement is
flexion/extension, thereby suggesting the first rotation in sequence should be X (or whichever
letter/symbol represents flexion/extension). This allows for avoidance of Gimble lock as it is
less likely that the remaining two rotations (ab/adduction and internal/external rotation) will
reach the 90° mark. The secondary rotation should likely be whichever the next plane of motion
of interest is. However, in the upper extremity, this is not always the case as many of the
rotations occur outside of the sagittal plane, in which rotations approach Gimble lock values.
Movements such as a golf swing, baseball throw often involve secondary rotations which
approach 90°, thereby making an approach to the singularity possible should the correct
rotational sequence not be selected. In a comparison of Cardan and Euler rotation sequences in
the upper extremity in examining for the presence of Gimble lock, there were instances of both
types of rotation where it was present and instances of both where it was not present (102). This
illustrates the necessity of the Euler rotation sequences containing a repeated axis as it is
mathematically sounder for certain movements in the upper extremity. An analysis of the
desired movement is necessary for determining the most appropriate rotational sequence. Euler
30
may be better suited for the upper extremity, but Cardan seems to be better suited for the lower
extremity where the excessive secondary rotations are seldom present.
In using the external GRF method, it was found to be highly sensitive to the reference
frame of choice, with altered magnitudes and profiles sometimes occurring (97). The GRF is
used with the free bodies of the foot and shank to calculate joint moments about the knee joint
center which can be referenced as the midpoint of the transepicondylar axis. However, when
combined with skin marker data, there has been a shift in the knee joint center during the first
40% of stance, leading to a reduced knee extensor moment (as high as 12.3%) due to the
posterior shift of the knee joint center (85).
Global coordinate systems (GCS) and local coordinate systems (LCS) are additional
methods which can be used. The GCS references motion of a segment within the global or
laboratory environment, essentially defining the motion within the entire calibrated area in which
the data is being collected. A set of axes is defined from a point of origin within the lab space.
An LCS is defined by the marker data on the segments and utilized in relation to an additional
segment. An orthogonalization process can be used to define the axes for the LCS. Similar to
the JCS, a set of markers can be used to define an initial axis. A perpendicular line is then drawn
towards an additional marker or an already defined joint center or landmark of some sort. A
cross product of the two existing axes then defines the third. This can be done in a sagittal,
transverse, frontal order as part of the Gram-Schmidt process (59). Rotations can then be
calculated for a segment with respect to the axes defined in the GCS or LCS, which should
theoretically be the same when properly expressed, however this is not always the case with data
reported between the differing coordinate systems.
31
In comparing GCS to LCS expressions, there were several differences with respect to the
lower limbs at different parts of the stance phase when reporting joint moments. For example, a
comparison of global versus local coordinate systems showed an over estimation of ankle
moments in the frontal plane at mid stance and the transverse plane at 40%, 60%, and 80% of
stance while underestimating the sagittal moments at 20% and 40% of stance. The knee
moments were overestimated in the sagittal plane at 20%, 40%, and 100% of stance and the
transverse plane at 20% and 100% of stance but underestimated in the frontal plane for the entire
stance phase and the transverse plane for 0%, 40%, 60%, and 80% of stance. Finally, the hip
was underestimated for all of the frontal plane and 20% and 80% of the sagittal plane, while
following the same transverse pattern as the knee (59).
Research has been performed to examine the correlations of joint moment calculations in
the GCS, LCS, and JCS. It showed that there were the lowest correlations in the transverse
plane. The sagittal plane showed strong correlations while the frontal had moderate (98).
Despite the correlation values though, magnitudes can still present as different which requires
caution with interpretation. During a side step maneuver, the sagittal knee showed high
correlations but the GCS system reported lower values than the LCS and JCS (98). In general,
sagittal profiles have little to negligible effects when using global or local frames of reference.
There is some discrepancy with the frontal plane though as while some report moderate
correlations, others have seen little influence with respect to the coordinate system (97). The
transverse plane seems to have a general agreement with errors between the different coordinate
systems, rendering these values ones to be interpreted with caution as the coordinate system
plays a big difference in the magnitudes and profiles, with the opposite profiles sometimes being
present (97, 98, 105). It may be of benefit for researchers to consider the motive of their
32
research prior to choosing a coordinate system as it has been suggested that using a GCS is
beneficial for examining the contribution of a joint to movement in a certain direction while
using an LCS is better for interpreting loading at a joint (98). As discussed earlier, STA is also a
factor in this. When examining the STA of the shank in different coordinate systems, the tibial
coordinate system was shown to be less affected that the femur, leading to the suggestion of
using the tibial coordinate system for calculations because of the decreased sensitivity (83). This
has the advantage of being an orthogonal coordinate system, but has the disadvantage of not
providing joint moments in the manner to which the research community has come to understand
joint moments. For example, a knee flexion moment in the tibial coordinate system does not
entirely correspond to the extension moment generated by the quadriceps muscle if the knee has
some simultaneous transverse rotation (98).
For the LCS, the development of the initial axis is an important consideration as there are
multiple methods by which to establish it. In a comparison of three different axes, differences
were evident in the calculations after the axial establishment. There are three common methods
for establishing the axis. In the knee, for example, an axis could be created at which flexion and
extension is perceived to occur by the investigator. This can be achieved through a knee
alignment device, however misalignment of it does present the potential for cross-talk errors
given that these are essentially planes of finite rotation with a mathematical interaction being
applied to them (65). Second, a line could be created based on anatomical landmarks, such as
between the two epicondyles, thereby creating the flexion/extension axis. For these two
methods, the other axis can be created through perpendicular lines to additional landmarks and
cross products, as described earlier. These are the two most commonly used methods, but a
dynamic method can be implemented based on an optimization procedure in which the
33
longitudinal femoral anatomical axis is defined and then a mediolateral axis is rotated about that
axis by a given degree value, designed to minimize the frontal plane profile variance (106). This
dynamic method has shown higher repeatability than the first two methods described. The
frontal plane inconsistency, which as mentioned earlier can be problematic, was lowest in the
dynamic method (106). This is likely due to the optimization procedure but it is significant
nonetheless as it was also shown to minimize the joint angle cross-talk at the knee. This method
has an assumption of minimal frontal plane knee movement, which appears to be acceptable for
reduced knee flexion motions (under 90°) which is accurate for many dynamic movements (106).
This method may be inappropriate for certain pathological populations where knee laxity in the
frontal plane is present, such as knee replacement patients or ACL-injury patients. A
hierarchical model has also been employed for calculations, but the primary issue with it is the
trickle-down effect for the errors in which the errors compound onto subsequent joints (65).
Ultimately the use of error estimates with clinical data can be beneficial as it increases
the objectivity of the data interpretation. This allows a focus on deviations which exceed the
acceptable level of experimental uncertainty (65). It has been suggested that all of these varying
methods of calculating kinematics and kinetics are valid, but that the decision should be carefully
evaluated for the population it is being applied to because the information is not necessarily
directly comparable. The JCS has been suggested as the best option for a standardized system as
it best represents what a joint moment actually is (105).
The final decision for what marker system to use, what calculation reference frame to
use, and axial creation method to use needs to be evaluated on an individual basis. This should
be based on the population of interest, how previous errors with certain styles may affect the data
desired to be collected, and whether or not the data will be comparable to other research. The
34
direct comparison between data sets is not always possible due to the fundamental differences in
methodology. This renders the need for caution when interpreting data and comparing to other
data sets as the results may not actually be as clear as initially seen. Independent analysis for the
errors and issues at hand must be taken into consideration before any comparisons are made and
definitive conclusions are drawn. Failure to do so may result in inconsistent or inaccurate
information being presented, which may have clinical implications that could prove detrimental
to the population in question. Great care must be taken in making these choices prior to the
execution of the project.
Biomechanics, Strength, and Balance of Total Knee Replacement Patients
Osteoarthritis (OA) is a disease characterized by degradation of cartilage within a joint.
According to the Center for Disease Control, approximately 22.7% of adults in the United States
(52.5 million) have a form of doctor-diagnosed OA (107). OA can occur in any joint and knee
OA is the one of the most common forms. It will afflict approximately 25% of the population by
the year 2030 (1). OA is a progressive disease where the wearing down of the cartilage can
eventually lead to bone on bone contact. As the disease nears the end-stage, a total knee
replacement (TKR) is one of the treatment options. In 2011, there were over 700,000 TKR
operations performed in the U.S. (2) with over 3.5 million per year projected by the year 2030
(3). With each replacement, there is a significant financial implication associated with the
procedure. In the state of Wisconsin, for example, the average TKR procedure costs
approximately $19,169 (108). It is projected that the total costs of TKRs by 2030, based on
projections, will exceed a value of $67 billion. In addition, the occurrence of TKR procedures is
increasing in patients under the age of 60 (4). Ultimately this means that people who undergo
TKR procedures will have longer lives to live with the replacement in their joint, making the
35
longevity of the replacements important as well. While restoration of function is typically the
most important factor in considering a TKR procedure, the addition of the high cost make this
an important issue worth examining in an attempt to make the outcomes as optimal as possible
given the significant financial investment.
The primary goals of a TKR procedure include a reduction in pain, improved knee joint
range of motion (ROM), improved knee joint alignment (as a malalignment often contributes to
the degeneration of cartilage), and a restored ability to perform activities of daily living (ADL),
with some patients having a desire to return to more advanced physical activities such as tennis,
cycling, golf, and swimming (5). Many of the common survey tools filled out by the TKR
population examine the patients’ desire to return to advanced activities beyond simple ADLs,
thus indicating a hope of many patients to resume advanced activities post-operation. Overall,
the operations are generally considered successful for the majority of patients. Reductions in
pain and improvement in ROM have been commonly reported (6-9), with accompanying
satisfaction rates ranging from 81-97% (10, 11). However, despite high satisfaction rates, this
still leaves a significant percentage of patients dissatisfied with the outcome of the procedure.
Many patients still report post-operative pain (12) and functional limitations (13), often resulting
in decreased performance on clinical tests (such as the timed up and go, six minute walk test,
and sit to stand test) compared to healthy controls (14, 15). These clinical tests are frequently
used as a means of assessing the success of an operation due to their ability to determine whether
a patient has restored function of the replaced joint. However, the dissatisfaction of TKR
patients is not sufficiently explained by the clinical tests and survey data.
Biomechanics data is often utilized to examine TKR gait patterns as a means of assessing
the progress of the patients, as there are known gait profiles which contribute to the exacerbation
36
of knee OA which leads to TKR procedures. Researchers have studied everything from simple
to complex variables such as gait velocities (20, 21), sagittal plane knee ROM (20, 22), and
frontal plane knee moments (20, 21) during overground walking as a means of searching for
detrimental movement patterns. Additionally, researchers have examined more demanding
movements such as stair climbing since it is a common activity in daily life for both older and
younger people (31). Stair climbing has been shown to be a difficult task for people with knee
OA (33), the same people who may become candidates for a TKR procedure. Current survey
tools such as the Western Ontario and McMasters Universities Arthritis Index (WOMAC),
Forgotten Joint Score (FJS), and Knee Society Knee Scoring System (KSS) all use stair climbing
as a means of assessing function following TKR and for gauging improvement (5, 34, 35).
Advanced physical activities, such as stair negotiation, may highlight movement deficiencies for
patients who may or may not show the same deficiencies on less demanding tasks such as level
walking (29, 109).
Most of the previously mentioned studies examined the TKR population as a whole and
did not differentiate between the satisfied and dissatisfied populations. There is a lack of data
regarding the biomechanical profiles of these two subgroups of the TKR population. There have
been minor explorations in survey-based data and how certain activities related to patient
satisfaction with the TKR procedure (110), how patient expectations were not met (111), and the
experienced changes in pain post-operatively (112).
TKR Patient Biomechanics of Level Walking and Stair Climbing
When examining biomechanics data, there are different methods utilized to provide a
baseline comparison to the most recent physical state of the TKR patients. These comparisons
include pre to post-operative physical states within the TKR patients as a means of assessing
37
progress post-operatively (27, 53, 113), between TKR patients and healthy control subjects (20,
22, 23, 38, 114), and post-operative compared to both pre-operative levels and healthy controls
(15, 24-26, 54). Gait is a common measurement protocol as it is the most commonly utilized
activity for the general population.
Stair climbing has also been studied as it is a commonly performed activity in daily life
for many people. More importantly, stair climbing represents a more physically demanding task
as it requires, even at a slow pace, twice the metabolic expenditure of slow walking (115), and
subsequent increased demands on the involved muscles and joints (29). This task allows
researchers to examine a task with increased demands as a way of assessing advanced physical
function for the TKR patients. Given potential limitations for the TKR population both pre-
operatively and post-operatively, these differences in the tasks are significant. Climbing stairs
has been reported as one of the top five most difficult tasks for people over the age of 60 (32).
This task is increasingly difficult for the knee OA population (33), the same people who are
candidates for TKR procedures, as they often have increased pain and functional limitations in
advanced physical tasks. Additionally, stairs are an infrequently assessed task in most surveys
utilized on the TKR population (5, 34, 35). While many people are able to live in homes without
stairs, they may eventually encounter them in a public realm where the alternative to the stairs
may not be a viable option, thereby highlighting their physical limitation. Within both walking
and stair climbing, there are many variables which play a part in the analysis of an individual’s
gait profile. These include spatio-temporal variables, ground reaction forces (GRF), kinematics,
and kinetics, which can be further subdivided into specific planar assessments.
38
Spatio-Temporal Gait Variables
Spatio-temporal variables are commonly assessed in both laboratory and non-laboratory
settings (such as physical therapy clinics or hospitals) because the data are easy to collect with
minimal equipment required. Often times, a simple stop watch and measuring tape allow for
recording of this information. This information includes gait velocity, stride length/frequency,
and step length/frequency. These values are easily tracked in a longitudinal fashion as a way of
assessing improvements in function over time.
Gait velocity is a simple tool for measurement as a fixed distance can be measured and
the patient can be timed with a stop watch to assess his/her velocity in covering the known
distance. Pre-operative gait velocities for TKR patients have been reported to range from an
average of 0.89 m/s (24) to 1.13 m/s (25). These speeds often relate to the severity of the knee
OA progression as well as the pain levels associated with the disease at the time of measurement
as pain can limit functional ability and cause the patient to reduce speed as a way to control the
pain. Post-operative velocities have been shown to increase compared to pre-operative levels
(24, 27, 53, 113) with increases ranging from 0.1 m/s to 0.17 m/s at one year post-operative (27,
53). The magnitude of the increase is dependent on the pre-operative value and based on the
overall functional increases of the patients. However, these results are not always consistent as
some researchers have reported no increase in gait velocity (25) or reduced gait velocities at two
months post-operative (15). This may suggest that improvements are recovery and/or time
dependent as not all variables have been shown to increase as early as two months, with
improvements still manifesting more than 12-months post-operatively. When a patient
undergoes a bilateral TKR, the gait velocity may be further impacted. Decreases in gait velocity
have been reported as long as 8 years post-operatively (36).
39
Improvements appear to be continuous throughout the first year post-operatively (27).
Different variables reach peak, or “normal”, values more than 12-months after the operation
while some reach them much earlier. Initial declines in function are common post-operatively
(15). It may take as long as one year before the operative leg reaches the same level as the non-
operative limb. Asymmetry has been reported at 3 months post-operative but disappearing by 12
months (46).
While improvement is regularly shown for TKR patients in comparison to themselves, it
is important to note that these levels of improvement may not reach those of healthy control
subjects (116). Gait velocity differences have been reported to be decreased between 0.12 m/s to
0.6 m/s for TKR patients when compared to healthy controls (15, 38). This may be reflective of
time from surgery as the magnitude of difference was larger at 2-months post-operatively (0.7
m/s for TKR and 1.3 m/s for healthy) and the smaller magnitude was at 1 year post-operatively
(1.31 m/s for TKR and 1.43 m/s for healthy), implying that recovery and restoration of function
takes time (15, 38). This result is not always consistent though as some studies have reported no
gait velocity differences between the two groups at 12-18 months post-operatively, with the TKR
patients retaining their velocity at 46 months post-operatively (22, 26). The inconsistent results
may suggest that additional studies are needed in order to provide a definitive conclusion on gait
velocity improvements as to whether a return to the levels of healthy controls is possible. It does
appear to be consistent that an improvement from pre-operative velocity is possible.
Stride length increases have accompanied the velocity increase when comparing pre-
operative and post-operative values. Mean increases of 0.04 to 0.11m have been shown (24, 53,
113). As with gait velocity, initial declines can be expected after the operation, with reductions
in stride length being reported at a two month follow-up period (15). Although one study found
40
no difference in stride length between pre-operative levels and post-operative levels at 12 months
after surgery (25). This may suggest that the population of TKR patients used here were higher
functioning prior to the operation than some other TKR populations. In comparison to healthy
controls, the results are mixed. Several studies have found reduced stride length for TKR
patients compared to healthy controls (15, 20, 23-25, 38) while two other studies have reported
no difference between the two groups (22, 26). Step frequency (how often steps are taken) has
been shown to be greater in controls as well (15, 20, 23, 25) with the same two studies who
showed no difference in stride length showing no difference in step frequency (22, 26). Peak
differences in step frequency reached as high as a 25 steps/minute reduction for the TKR group
at 2 months post-operation (110 steps/minute versus 85 steps/minute) although the differences
did reduce as time progressed (121 steps/minute versus 115 steps/minute) (15, 25).
Improvements did exist for the TKR population after the operation but did not reach the levels of
controls.
With stair climbing, there are slight differences as step length is a little more complex to
assess given the fixed size of the stair, thereby making alterations tough to achieve. In 2003,
Nadeau et al showed reduced stride lengths and frequency on stairs and a decreased forward
velocity coupled with a lengthened swing phase compared to overground walking for adults over
the age of 40 (29). These alterations are an indication of the complexity of the activity of stair
negotiation in comparison to overground walking. Additionally, healthy individuals have been
found to have a slower velocity when ascending the stairs (0.49 m/s) compared to descending the
stairs (0.56 m/s), further suggesting the complexity of a stair ascending task (117). As a whole,
spatio-temporal variables are less frequently reported for stair negotiation compared to
overground walking.
41
Ground Reaction Force
Ground reaction force (GRF) is the force applied to the body from the ground during the
stance phase of a motion. It is often used to assess external loading to the body during dynamic
tasks, with the reporting often done in a normalized version (usually a percentage of the
individual’s body weight) of the absolute value in Newton. Additionally, the loading rates are
often reported as a descriptive measure of how quickly the external force is applied to the body.
There have been similar vertical GRF values and loading rates shown between the operated and
non-operated limb for unilateral TKR populations (114). The results of this study did span a
wide amount of times post-operatively, with subject times from surgery ranging from 4 to 96
months.
Pre-operatively, research has shown that during walking the non-operated limb has had
higher peak vertical GRF values (1.06BW) compared to the operated limb (1.03BW), which has
a resultant effect size of 0.297. Post-operatively, this same group saw increased loading to both
the non-operated and operated limb (1.10 and 1.06 BW, respectively, effect size=0.244). These
increases were significantly different from pre-operation to post-operation as well as from each
other. The loading rates for these same subjects increased post-operatively, but the two limbs
were not different from each other after the operation (113). Pre-operative favoritism of the
healthy leg is to be expected as a way of taking some of the load off of the diseased limb, likely
due to pain. However, with continued asymmetry between the limbs, it should come as no
surprise that approximately 40% of the patients who have unilateral TKR procedures have their
contralateral limb replaced within a ten year time period (118). Other researchers have shown a
return to vertical GRF symmetry by 12 months after the procedure despite significant differences
occurring at 3 months after the procedure (46).
42
The VGRF profile is usually a bimodal curve. The first peak represents weight
acceptance and the second curve is the force applied during push-off (113). The values of the
first peak VGRF are greatest during stair descent compared to stair ascent or level walking (117,
119) with values reaching nearly 1.5 BW (119). The second peak has been reported higher in
stair ascent compared to descent and overground walking (117). This is to be expected given the
requirement of a higher propulsive peak in order to make the vertical ascent to reach the next
step in the stair case. Additional research has shown walking as having higher weight
acceptance peak VGRF (1.13BW) compared to stair ascent (1.04BW) with no push-off
differences between the two conditions (120). Bone on bone reaction forces of 4.25 BW have
been reported when walking up stairs (121). In a more recent study, joint reaction forces (as
calculated through musculoskeletal modeling) has shown early stance peak compressive forces
reaching 2.76BW and late stance at 3.89BW during stair ascent (122). This number is crucial for
the TKR population as altered loading is a part of the contribution to end-stage OA leading to a
TKR procedure. Controlling the forces is crucial for longevity of the joint and maintaining joint
health.
The highest peak compressive forces are experienced during stair descent in comparison
to other ADLs, with knee forces reaching levels of 1.23 BW and hip forces reaching 1.1 BW,
both of which were approximately 0.2 BW higher than walking values as assessed through
inverse dynamics (123). Patello-femoral contact force has been modeled for both stair ascent
and walking, resulting in a contact force 8 times higher during stair ascent in comparison to
walking (33). While stairs may be avoidable for some patients, ideally there would be a desire to
return to normal function after the TKR procedure, which would include stair negotiation when
necessary. Given the increased ground reaction and joint reaction force levels in comparison to
43
walking, it becomes more important for successful control of those forces as they are applied to
the body.
Sagittal Plane Kinematics
Level Walking
Studies examining TKR patients have looked at sagittal plane kinematic variables
including ROM, flexion contact angles, and maximum knee flexion during both stance and
swing phases of gait. One of the goals of the TKR procedure is a restoration of sagittal knee
ROM so it is frequently assessed by the surgeon and physical therapists. It functions as a
measure of determining success of the procedure. Static ROM is a frequently used measurement
tool by surgeons and physical therapists, but biomechanical analysis can add to this assessment
by examining more dynamic aspects of movements for patients as the ultimate goal is a return to
normal independent functioning.
Knee flexion angle at contact during walking has been reported as similar between TKR
patients and control subjects by several studies (20, 23, 25) but most TKR research has reported
reductions in maximum knee flexion during the stance phase compared to the control subjects by
an average of 6° (15, 22-24). A reduced gait velocity may be related to the reduced active knee
flexion ROM. Only two studies have reported similar maximum flexion angles between TKR
patients and healthy controls (25, 26). This reduction may carry on through the first year after
the procedure. A reduced dependence on the knee in the operated limb has been shown at 12
months post-operatively compared to the non-operated limb and healthy controls (46). There is
typically an increase in maximum flexion angles post-operatively, but one study has shown no
difference in maximum flexion angle or flexion ROM in comparing pre- and post-operative
levels although both values were still reduced in comparison to healthy controls (116). It is
44
important to note that this study examined differences at only three-months post-operatively, and
that the dynamic ROM may not have fully recovered yet. Additionally, reduced knee extension
during mid-stance has been reported for TKR patients compared to healthy controls, with
absolute value differences of 10° pre-operatively and 7° post-operatively (25), suggesting no
difference between the two groups for knee contact angle.
During the swing phase of a level walking gait cycle, some reports of reduced knee
flexion have shown a reduction of approximately 10° for TKR patients compared to healthy
controls, with the TKR patients producing 50° of flexion compared to 60° by the controls (20,
22, 23). Yet other studies have reported no difference between the two groups for swing phase
knee flexion (25, 26). Despite this, TKR patients still have less knee flexion for both the stance
and swing phases. This leads to a reduction in ROM as it is typically measured as the difference
between contact angle and the maximum flexion value achieved during stance or the toe-off
angle and the maximum flexion angle during swing. The decreased flexion ROMs have had
reported values between 8-15° (20, 116). The reduced flexion by the TKR population has been
speculated to be associated with a quadriceps avoidance pattern which the TKR population
employs as a method of unloading the knee joint following their operation (124). Furthermore,
the differences between these studies where no differences were found (which seems to be in the
minority) may indicate a higher level of functioning for the TKR sample population used. Most
TKR studies highlight kinematic as well as spatio-temporal parameter differences, thereby
attempting to link the two variable types. Similar gait velocities have been shown at the same
time as similar sagittal plane kinematics when comparing TKR patients and healthy controls
(26), further highlighting this link.
45
It is important to note that time points from surgery make a difference in the results. Pre-
operative and post-operative comparisons highlight recovery over time, but not all post-operative
results are directly comparable because of the time differences. At 2 months post-operation, a
difference of 9° in peak knee flexion during the stance phase of level walking can be seen in
post-operative values compared to pre-operative (15). By 6 months, this value has been shown
to decrease to 3° (24). By one year, peak knee flexion has been shown to increase beyond pre-
operative levels (53). Although some studies have shown no differences in pre to post-operative
differences in peak knee flexion at 3 months post-op, there were also no changes in gait velocity
(116), which therefore should require less knee flexion so this result is to be expected.
Stair Ascent and Descent
During stair negotiation, the sagittal kinematics are different than walking. The ankle is
in increased dorsiflexion during stair ascent compared to walking as the dorsiflexion is required
at the onset of stance during stair ascent (29, 117) whereas in walking the dorsiflexion occurs
later in stance. Simultaneously, the knee is in increased flexion at contact (65°) compared to
walking (1°), coupled with much higher values of maximum flexion (93° in stair ascent
compared to 67° in walking), resulting in an increased flexion ROM when measured from
maximum flexion to maximum extension (29, 117, 125). Healthy adults have an increased ROM
during stair ascent compared to walking (125), therefore making the increased ROM for TKR
patients during stair negotiation expected. Compared to healthy adults, TKR patients have a
reduced knee flexion value at contact of between 8.7° and 17.7° (24, 39-41, 126). There is an
accompanying reduction in swing phase maximum flexion angles of 11-15° during stair ascent
(40-42, 126). One study found no significant difference in maximum knee flexion during swing
phase for TKR patients, however they reported absolute mean differences of close to 10°
46
compared to healthy controls (39). Knee flexion ROM has been shown to be reduced compared
to healthy controls in both stair ascent (22, 36, 37, 39-41) and stair descent (22, 36).
The differences in TKR patients may be related to the prosthetic design as when
comparing to healthy controls, one study found that TKR patients with a non-resurfaced patella
had 17° reductions in flexion ROM (39) while another study found patients with a mobile
bearing design only had reductions of 10° (41). However, two additional studies found no
difference in flexion ROM when comparing TKR to healthy controls (22) and when comparing
the replaced to the non-replaced limb on stair ascent (37). Two other studies also found no
differences when making the same comparisons on stair descent for healthy controls (40) and to
the non-replaced limb (37). Unfortunately, recent studies have made no comparisons during stair
descent in comparison to healthy controls or to pre-operative levels (127, 128), making
improvement comparisons difficult as the body of research on the topic is currently very small.
Sagittal Plane Kinetics
Level Walking
Kinetics are often measured along with kinematics as another way of assessing the
biomechanical gait profile of a subject. Kinetic data provide information on joint loading
through the analysis of joint moments. The most commonly utilized kinetic measurement is the
joint moment which is defined as a linear force value multiplied by the perpendicular distance at
which that force acts from the axis of rotation (moment arm), causing rotary motion at the joint.
These moments are reported as either internal or external moments, depending on the calculation
convention used by researchers. The joint moments are frequently normalized to either body
mass (Nm/kg) or bodyweight and height (%BW*height).
47
While internal and external moments are often used interchangeably for comparative
purposes, this is not always an exact comparison as the moments effectively represent two
different things. They are similar in nature but external moments represent effects by GRF while
the internal moments represent the moment by internal muscle forces. These differences are
important to highlight in making comparisons between the two measurement conventions.
During walking, peak external knee extension moments (often used interchangeably with
internal flexion moments) have been reported to be reduced in TKR patients compared to healthy
controls, with group differences of 0.16 Nm/kg and 2.2%BW*height (15, 20, 25, 129). When
examining the peak external flexion moments (occurring during early stance), healthy controls
were reported to have significantly higher moment values (0.30 Nm/kg) compared to TKR
patients (0.22 Nm/kg) during walking tasks (26). Internal extension moments at weight
acceptance have been shown to be reduced by 1.9%BW*height for TKR patients compared to
healthy controls (24). As with some of the kinematic variables, these differences can be time
dependent. At 2 months post-operation, the TKR patients showed a maximum internal knee
extension moment of 0.18 Nm/kg, which was significantly lower than the healthy control value
of 0.34 Nm/kg while the internal flexion moments were also lower for the TKR patients
compared to the healthy controls (15). The reductions have been reported as long as 46 months
post-operation (22), although the magnitude of difference appears to shrink over time.
Conversely, there are studies which report the maximum external flexion moment in TKR
patients as similar to those of healthy controls (20, 25). This lack of difference is speculated to
occur possibly as a result of the TKR population retaining their pre-operative gait which was
matched with the healthy controls for maximum external flexion moments. The presence of a
decreased external extension moment suggests a quadriceps avoidance pattern, which may have
48
been a movement trait retained from prior to the operation. It is suggested that a therapeutic
program to rectify this may be necessary (25).
Post-operatively, TKR patients have seen a more bimodal pattern curve for sagittal plane
joint moments compared to pre-operatively (53). This pattern is more characteristic of a normal
loading pattern seen in healthy individuals, suggesting a return of the TKR patients to a more
normal kinematic pattern. There have been reports of no differences in maximum knee
extension moments (likely internal moment although they failed to mention if it was internal or
external joint moments) from pre- to post-operation, with a decrease in the knee flexion moment
of 0.20 Nm/kg at two months post-operation (15). This may have been too early of a time period
for a measurement for assessment for return to normal levels of healthy controls as has been seen
with other variables. Two tests which reported on subjects at periods of less than 6 months still
showed differences between TKR patients and control subjects (15, 24). At one year post-
operation, maximum external extension moments in early stance have been reported as being
similar between pre- and post-operation time points, but maximum external flexion moment at
early stance was increased for TKR patients by 0.93%BW*height (25), further indicating that
recovery is definitely not complete by two months as the evidence still indicates changes at one
year post-operation. An additional study reported within subject differences for internal
extension moments at the weight acceptance phase decreased post-operatively by
0.9%BW*height (24).
An additional study has reported that there was no asymmetry between replaced and non-
replaced limbs for sagittal moment patterns during walking (130). This is not always the case
though and it may indicate a change in joint dependence as the contribution of the knee extension
moment to lower extremity support has been shown to decrease at 12 months compared to three
49
months for the operated limb, but not for the non-operated limb, with both still being lower than
healthy controls (46). This may indicate the quadriceps avoidance pattern previously mentioned
as a means of unloading the knee joint and thus further relying on the non-operated limb.
Stair Ascent and Descent
When comparing stair ascent to overground walking, peak external knee extension
moments are greater in ascent than in walking, with value differences ranging from 0.18 Nm/kg
to 0.52 Nm/kg between the two tasks (29, 117, 131). For the peak external knee flexion moment
(often considered more important for examination than the external knee extension moment),
values reported as three times greater have been seen during stair ascent compared to walking
(131), indicating the advanced difficulty of the task and the required muscular effort in order to
complete the task. Other studies have shown the values of flexion moments to be 11.9
Nm/%BW during stair ascent which were significantly greater than 7.4 Nm/%BW during level
walking (123). Simultaneous maximum external hip flexion moments are also higher during
stair ascent (0.76 Nm/kg) compared to level walking (0.52 Nm/kg) and approximately 1.5 times
higher during stair descent compared to walking (29, 131).
Maximum internal knee extension moments have been shown to be reduced in TKR
patients (2.08-3.3%BW*height) compared to healthy controls (5.10-6.50%BW*height), with the
reductions potentially reaching in excess of 50% of %BW*height (24, 39-41). Despite what
appears to be substantial differences, there are other studies that have reported no differences
between the two groups, although they failed to report velocity of the stair ascent for both groups
(22, 36, 42), which as indicated earlier, may play a role in the required loading of the body
during gait. Reductions in maximum knee extension moments have been shown to occur during
simultaneous reductions in gait velocity for TKR patients (24, 39-41). A review study published
50
in 2014 showed that at the time, four studies had measured velocity and knee moments
simultaneously and all four showed decreased moments with reduced gait velocity during stair
ascent (132). These reductions in velocity (especially during stair ascent) may be related to
strength deficits of the TKR patients thereby reducing their ability to navigate the stairs more
quickly or related to pain levels experienced at higher velocities, thereby requiring them to slow
down to alleviate the pain. Based on the differences, a reporting of velocities is important as the
speed differences may be related to other kinetic differences. For example, a reduced peak knee
extension moment was seen in the replaced limb of TKR patients during stair ascent compared to
the non-replaced limb and healthy controls when no significant difference between velocity was
reported for the two groups, although the absolute mean velocity was lower for the TKR patients
(19). In this instance, velocity did not appear to play a role, but based on previous differences, it
may still be a significant variable and should be reported either as a means of disqualifying it or
highlighting it as a potential confounding factor.
Stair descent is, again, much less studied. One report has shown differences in the
maximum external knee flexion moment between healthy controls and TKR patients, with
reported values of 16.3%BW*leg length (LL) and 13.1%BW*LL, respectively (22). More recent
studies on stair descent have reported no differences between the two groups (36, 40, 127). The
conflicting results may indicate a need for more research to be done on stair descent as the
volume of research in comparison to overground walking and stair ascent is much smaller.
Frontal Plane Kinematics
Level Walking
It has become well understood that the progression and severity of medial compartment
knee OA is directly affected by the frontal plane mechanics an individual employs during gait.
51
The internal abduction moment has been suggested as a surrogate measure for loading to the
medial compartment of the knee joint and has been linked to the progression of medial
compartment knee OA (133-136). Additionally, a correction of the frontal plane knee alignment
is one of the goals of the TKR procedure.
After the correction of frontal plane knee alignment, TKR patients have shown a
reduction in the peak knee adduction angle during gait, with levels returning to close to those of
healthy control subjects. TKR patients have shown peak knee adduction during the stance phase
of 4.1° while healthy controls show 3.9° peak angles (23, 54). An additional study has shown
that by 6 months post-operation, the absolute values of the peak knee adduction angles for TKR
patients have returned to levels of 3.6° (27), similar to healthy control levels. These values were
not directly compared to healthy controls, but the absolute magnitude is similar to other reported
values for healthy controls. While not directly reported in this study, Orishimo et al (2012)
displayed an ensemble average curve for knee adduction angles which showed different initial
contact angles for TKR patients after surgery. Pre-operatively, initial contact appears to be
around 3° of adduction, leading to a peak adduction of 9.7°, while post-operatively initial contact
is between 0-1° of abduction, leading to a peak angle of 3.6° (27), suggesting a decreased ROM.
The values for peak knee adduction angles are directly related to frontal plane knee
alignment, and given that correction of the alignment is one of the TKR goals, surgeries have
been shown to successfully achieve this goal (27). The correction of the alignment is a static
measure, but it appears to have dynamic implications given the reductions seen in the peak knee
adduction angles during dynamic tasks. A recent study on healthy individuals with different
static knee alignments has shown differences in peak frontal-plane knee angles and knee joint
moments, although no differences in frontal-plane ROM were present between the different
52
static alignment groups, further supporting different contact angles for the different groups (137).
Furthermore, the operated limb has been shown to have reduced peak knee adduction angles
(1.8°) in comparison to the non-operated limb (4.3°), and sometimes reported as even being
lower than healthy controls (2.4°), although this was on a non-significant level (114).
By fixing the alignment of the knee, less adduction has also been found at the time of the
peak knee adduction moment in the operated limb (0.9°) compared to the non-operated limb
(3.28°). Neither limb was considered different from healthy controls (2.5°), only from each
other, but this continues to highlight the improvement in frontal-plane mechanics for the
operated limb (38). At 6 months post-operation, the peak knee adduction angles have been
shown to significantly reduce from 9.7° to 3.6° (27), resulting in relatively quick adjustments to
the frontal kinematics in comparison to other measures which have been shown to take up to a
year to normalize. Reductions in peak addiction angle were also seen at the time of peak VGRF
when comparing pre-operation values to post-operation values, with decreases of 4° post-
operatively (54), further highlighting the alignment correction benefits.
Stair Ascent and Descent
In stair climbing, peak knee adduction angles significantly increase compared to walking
(10.4° and 4.6°, respectively), with simultaneous increases in peak ankle adduction, but no
differences in frontal plane hip kinematics (29). At an average of two years post-operatively,
there were no differences reported in peak knee adduction angles when comparing the operated-
limb, the non-operated limb, and healthy controls (19). This was coupled with reduced ankle
eversion angles for the operated-limb compared to the non-replaced and controls (19), suggesting
that the alignment correction may have impacted the frontal plane ankle angles as well.
53
The peak knee adduction angles were also coupled with an increased ROM on stair
ascent compared to walking. In one study, walking had a reported frontal-plane ROM of 10.6°
while stair ascent had 15.4° (29), although no statistics were reported with respect to this
variable. This is primarily the result of the significantly increased peak adduction angle. Other
studies have failed to report on frontal-plane ROM as well, although given the similar peak
adduction angles reported (19), it is possible that the ROM are similar.
Frontal Plane Joint Kinetics
Level Walking
The internal knee abduction moment (or external knee adduction moment; KAM) has
been linked to medial compartment knee joint loading, with higher loading suggested as being
due to higher moment values (133-136). There is often a bimodal pattern shown during gait for
this moment so comparisons are frequently made for both peak values. The first peak (loading)
is typically higher than the second peak value. This loading on the joint contributes to the
degeneration of cartilage for patients with knee OA and subsequently can contribute to the
degradation of the joint replacement, thereby making controlling the loading important. Given
the increased frequency of joint replacements and the decreasing age of the recipients, it is
important to enhance the longevity of the replacement, and therefore control the loading.
There are contrasting results present with respect to the KAM value and comparing TKR
patients to healthy controls. Smith et al. (26) found no significant differences between TKR
patients (0.39 Nm/kg) and healthy controls (0.46 Nm/kg), while Benedetti et al. (20) found
reduced first and second peak KAM values compared to healthy controls, with the first peak
being reduced by 1.4%BW*height and the second peak by 1.0%BW*height (20, 26). This point
can be further confounded by research which has shown non-operative limbs to have increased
54
KAM values, with values reaching 0.07 Nm/kg in one study (38),0.67%BW*Ht in an additional
study (138), and 0.012 fat free mass/height in another study (114) compared to the operated limb
(38, 114, 138). However no differences were reported between replaced limbs and healthy
controls in those studies, which attests to the improvements of faulty mechanics in the operated-
limb being partially corrected through the operation, suggesting a return to healthy KAM levels
post-operatively. However, a further point of consideration is what this may mean for the
contralateral limb. The contralateral limb may be a weak link eventually requiring the same
procedure due to the increased loading which likely led to the need for a TKR in the first place or
it may be adapting to maintain a symmetry with the replaced limb after the aforementioned
adaptations by the replaced limb.
At 6 months post-operation, reductions in the peak KAM values were evident with a
decrease to 84% of the pre-operative KAM levels (3.2%BW*height to 2.7%BW*height) (27).
At 1 year post-operation, however, the same subject group returned to an increased KAM level
(3.0%BW*height) which was not significantly different from pre-operation levels (27). There
may be some slight regression in progress occurring over time, which could theoretically be
related to any number of factors. This same group saw a non-significant increase in walking
speed from pre-operation (0.93 m/s) to 6-months post-operation (0.99 m/s) followed by a
significant increase at one year post-operation (1.03 m/s) (27). At an average of 2 years post-
operation, another TKR patient group showed increased peak internal abduction moments in
their replaced limb compared to healthy controls, but showed no difference between the replaced
and non-replaced limbs or the non-replaced and controls (19). There were no gait velocity
differences for these two groups so velocity was a non-factor in this assessment, however the
knee extension moment was lower in the replaced limb compared to the healthy controls (19). It
55
does illustrate that the TKR patients may undergo a compensatory transformation whereby the
load is transferred to another plane as a compensatory mechanism. The reduced knee extension
moment may have been moved to the knee abduction moment.
Different time points within the gait cycle are of interest to researchers. At 6 months
post-operation, when measuring the internal KAM at peak VGRF, TKR patients experienced a
reduction from 4.07%BW*height at pre-operation measurement to 3.01%BW*height at 6
months. Healthy subjects in the same study had a value of 2.7%BW*height, which was not
significantly different from the post-operation values of the TKR patients but was significantly
lower than the pre-operation values (28). This may illustrate that the time of measurement (time
within the gait cycle) may play a difference rather than the absolute peak values. The article fails
to mention whether the peak KAM and peak VGRF occurred simultaneously but did state they
looked at the first peak VGRF time point since it corresponds to a time when the body is in
single-limb support (28).
There seems to be a general consensus that the KAM values will reduce post-operatively
(at least temporarily) but they will not be reduced beyond the level of healthy controls, with the
exception of one study finding a reduction in the external KAM for TKR patients compared to
healthy controls, implying a significantly reduced medial compartment loading (20). Given the
similar frontal plane kinematics found post-operatively, most of these findings are to be
expected. Reduced peak knee adduction angles following surgery (27, 54) would likely decrease
the moment arm, thereby reducing the frontal plane moment experienced at the knee. Reduced
varus (adduction) ROM occurred as early as 6 weeks post-operatively with an accompanied
reduction in the first peak KAM on the operated limb (138).
56
Stair Ascent and Descent
In comparison to level ground walking, the first peak KAM has been shown to
significantly increase (0.61 Nm/kg compared to 0.78 Nm/kg) for healthy adults with the highest
KAM values found during stair descent (29, 123). Additionally, the greatest hip adduction
moments are also found during stair descent (8.4%BW*height) in comparison to other common
ADLs such as standing, walking, rising from a chair, and bending over (123). It has been
suggested that the hip abductor muscles are needed to elevate the pelvis so that the needed
clearance is available for the swing leg to move up to the next step (29). When comparing TKR
replaced limbs to healthy controls, increased internal hip abduction moments were shown in the
replaced limb, with no difference for the non-replaced limb compared to controls. This was
coupled with reduced ankle inversion moments for both loading and push off for the replaced
limb compared to the non-replaced and healthy controls, however the knee moment differences
were mixed, with no differences for extension push-off moments or loading response abduction
moments. The replaced limb had reduced knee extension moment during loading and increased
abduction moment during push-off (19). This suggests an increased dependence on the hip in the
replaced limb upon a return to normal knee function in the frontal plane.
Replacement design also appears to play a significant role in frontal plane kinetics.
Posterior stabilized designs have shown KAM values equal to or reduced compared to healthy
control subjects as have mobile bearing designs, with KAM values of 1.8%BW*height for the
replaced limb compared to 2.7%BW*height for healthy controls (40-42, 54). Designs with non-
resurfaced patellas, however, have shown increased peak external KAMs of 3.8%BW*height for
TKR limbs compared to 2.7%BW*height for healthy controls (39). There seems to be a general
consensus on a restoration of frontal plane knee moments for TKR patients to the levels of
57
healthy controls during stair ascent, part of which is likely based on the restoration of knee
alignment post-operation and the increased muscle strength. During stair descent, which is
significantly less studied, results from two different studies have shown no differences in peak
internal knee abduction moments between TKR patients and healthy controls (42) and no
differences in peak external knee adduction moments for the same group comparisons (40).
Ultimately, there needs to be extended research into the biomechanical profiles of the
dissatisfied population as this information may provide insight into the reasoning for the
dissatisfaction with the procedure. Currently there is a significant lack of research into the
biomechanical variables as they relate to patient satisfaction with the TKR procedure. It is
possible that there is a connection between movement restoration to the levels of healthy controls
and those patients who are satisfied with their procedure. However, to date, this investigation
has not been completed, warranting the research into potential movement abnormalities post-
operation which may contribute to dissatisfaction with the surgical outcomes.
TKR Patient Knee Strength
In order to have a properly functioning joint, sufficient muscle strength is required. TKR
patients, as with almost any operation, experience a disturbance in their strength levels post-
operatively, given the invasive nature of the operation. A return of strength levels are imperative
for patients to be able to return to normal functioning. Failure to acquire the needed strength
post-operatively can impair patients’ abilities to perform their normal ADLs as well as more
complex activities which they may desire to participate in, such as tennis, hiking, and others.
Quadriceps strength is the most commonly measured strength variable for TKR patients,
however, hamstring strength is important as well given its functions at the knee joint. The
58
predominant factor referenced when examining strength is the peak torque generated by a
muscle, either through isokinetic or isometric measurement.
Isokinetic Strength
Schroer et al. (45) showed a return to pre-operative levels for quadriceps peak torque by
three months post-operatively. By 6 months, they were 17% stronger and 30% stronger by one
year. The hamstrings for the same population showed strength increases of 10.5% by 3 months,
26.6% at 6 months, and 36.2% by one year. Additionally, there were no differences in
quadriceps to hamstrings strength ratio at any time point (45), suggesting that the operation
affects both muscle groups equally.
Early on in the rehabilitation process, significant reductions in muscle strength are
evident. At one month post-operatively, TKR patients have shown 60% reductions in quadriceps
strength compared to pre-operative levels (43). This is likely due to two factors: atrophy of the
muscle (which is associated with surgery due to inactivity) as well as likely a reduced voluntary
activation of the muscle, which functions as a protective mechanism for the TKR patient as
increased activity at the joint can lead to pain. The reduced strength level early in the
rehabilitation process is to be expected given the reduced physical function mentioned previously
at early time points. An additional group showed deficits in strength compared to pre-operative
levels with upwards of 40% reductions in strength for the knee extensors and 34% reductions for
the knee flexors at one month after the procedure but a return to pre-operative levels by 3 months
(44).
The operated limb and the non-operated limb are both affected by the TKR procedure
(likely due to reduced activity levels post-operatively). Both limbs have shown reductions in
strength at 6 months post-operatively in comparison to strength levels of healthy controls (52,
59
139). By 12 months post-operatively, the non-operated limb has shown recovery to healthy
control quadriceps strength levels but the operated limb still showed reduced strength levels
compared to controls while showing equal strength to the non-operated limb (46). The absolute
values of isometric quadriceps strength are highest for the controls (34.2 N/BMI), but
statistically the operated limb (24.6 N/BMI), while different from the healthy controls, was
similar to the non-operated limb (28.4 N/BMI), which was similar to the healthy controls. There
have been instances of strength deficits of 60% in the knee extensors and 11% in the knee flexors
occurring as much as 6 years post-operatively when compared to healthy controls (140).
At an average of 10 months post-operatively, the replaced limb compared to the non-
replaced limb showed deficits for both knee flexors and extensors (141). The knee extensors
showed a 27% peak torque deficit, 23% power deficit, and a 14% cross sectional area deficit.
The knee flexor deficits were smaller but still present. Peak torque showed a 13% deficit and
peak power showed a 19% deficit (141). These results however, are not consistent across the
research. The deficits are frequently different based on the measurement protocol. At 30 days
post-operatively, a group of bilateral TKR patients exhibited reduced peak knee extension torque
when testing at an isokinetic speed of 180°/s, with a return to pre-operative levels by 60 days
after the operation (142). However, at an isokinetic speed of 60°/s, there were still peak
extension torque deficits at 60 days after the operation (142).
In a comparison of posterior stabilized and posterior cruciate retaining designs to healthy
controls, the TKR populations showed reduced peak torque for both the quadriceps and
hamstrings when testing at 180°/s but not at 60°/s compared to controls (there were no
differences between the two replacement types though) at an average of 98 months post-
operatively (36). For the hamstrings, the posterior stabilizing showed a 55% peak torque
60
reduction while the posterior retaining showed a 69% reduction. For the quadriceps, the
posterior stabilized showed 36% reductions while the posterior retaining showed a 43%
reduction (36). The failure to achieve higher peak torque values at the higher velocity may
imply a decreased ability for high velocity contractions, but does suggest an ability to return to
strength at the slower velocities. It may be important to note that while the statistics do not show
differences, the absolute values are lower for the TKR groups. One additional study has shown
no difference at both speeds of 180°/s and 60°/s between cruciate retaining and cruciate
substituting replacement types (143). In this study, the quadriceps and hamstrings were
measured at the two aforementioned speeds for isokinetic strength and showed no differences for
peak torque between replacement designs (143), illustrating that there is likely no difference
between replacement types in terms of strength.
Strength has also been measured for the purposes of comparisons with other elements of
physical recovery. Increased levels of pre-operative strength in the quadriceps has been shown
to increase functional abilities at one-year post-operation. The probability of lower functional
abilities (as assessed by the Short Form 12 for physical function) when displaying poor levels of
pre-operative strength is 2.28:1 (144). This lends credence to the idea of doing pre-operative
strength training in an attempt to increase functional recovery. At one month post-operation,
pre-operative strength training showed significant improvement in a sit to stand task (145).
Additionally, the training led to reduced strength asymmetries between the replaced and non-
replaced limb, while the asymmetries persisted in the control group due to decreased quadriceps
strength in the operated limb and increased quadriceps strength in the non-operated limb (145).
However this is not always effective. A pre-operative strength training group showed no
difference at 3 months post-operation compared to a group with no pre-operation training, with
61
both groups returning to baseline strength levels (146). Additional testing took place at 6-weeks
post-operation, however, neither group had progressed to pre-operation levels (146), suggesting
that recovery to pre-operation levels occurs sometime between 6 weeks and 3 months. It is of
interest to note that in that study, 90% of the strength training group felt the pre-operative
training was beneficial, but based on the results the researchers were unable to conclude that the
training was necessary. The argument could be made that the mental wellness given by the
program may be just as important as the physical wellness. An aquatic therapy program had an
initial increase in quadriceps strength but by 180 days post-operative, the strength levels were
similar to regular ward therapy groups (147). However, the aquatic therapy did increase the
ROM at both 90 and 180 days post-operation (147). Peak torque values were also not shown to
be predictive of single limb static balance after surgery, however this was for a measurement at
11 days after surgery (51).
Knee power has been predictive of stair negotiation speed. Larger extension power
deficits in the operated-limb and low flexion power in the non-operated limb were shown to be
predictive of slower stair ascending and descending speeds (141). When pre-operative
quadriceps strength was added to a regression model with age, flexion ROM, and pain, the model
was able to significantly increase its ability to predict post-operative timed up and go test and
stair climb performance (43). Knee extensor strength also has a significant positive correlation
to gait speed (50).
Isometric Strength
In addition to the differences with testing speed for isokinetic testing, differences have
been shown between isokinetic and isometric testing procedures. Isometric tests at a knee angle
of 75° have been shown to have more pronounced differences for both flexion and extension
62
strength. When comparing the replaced and non-replaced limbs, the operated limb had 39%
flexion deficits and 29% extension deficits compared to the non-replaced limb (148). Isokinetic
differences were smaller with 7% reduction during 30°/s for flexion and 22% reduction during
120°/s for extension (148). Time periods between the tests are also of additional consideration
as the isometric flexion decreased at 6 months post-operation while the isokinetic flexion
increased at both 30°/s and 120°/s (148). This further illustrates the need to take methodological
considerations when making comparisons between different studies.
Ultimately, a return to strength levels is crucial for the successful return to a functional
life as sufficient muscle strength is necessary for more complex activities such as stair climbing,
hiking, or squatting. It is important to return the strength levels from prior to the operation of
both the quadriceps and hamstrings in order to maintain the strength ratio as well as to maintain
symmetry between both limbs. Deficits are likely to occur early in the rehabilitation process, but
the general consensus is that a return to pre-operative strength levels are possible for TKR
patients.
Currently, there is a lack of data with respect to strength and patient satisfaction levels.
While certain activities have been shown to be more difficult for dissatisfied patients, this has not
been studied on connection with knee flexor and extensor strength levels. There is a need for an
examination of strength levels with the dissatisfied population to compare their strength with
satisfied patients. This may help to examine the causes of dissatisfaction post-operation and
thereby potentially help to improve the satisfaction rates should an identifiable cause be present.
TKR Patient Balance Abilities
Static and dynamic balance have become another measure of success for TKR operations
as balance is necessary to avoid falls and consequent fractures, which can be detrimental to the
63
TKR (47). Studies examining balance have occurred, like with strength and kinematic/kinetic
studies, with measurements both pre-operatively and post-operatively, single leg versus double
leg, static versus dynamic, and eyes open versus eyes closed. Different systems have been used
including the Biodex Stability System (51), 3D motion capture (52), COP analysis on a force
platform (139), and the Nintendo Wii Balance Board (50). Variables of interest have included
anterior-posterior stability index, medial-lateral stability index, overall stability index, velocity,
and sway paths for the center of pressure (COP). Anterior-posterior, medial-lateral stability, and
overall stability indices are calculated based on degrees of tilt of the balance platform from a flat
horizontal plane. The indices are the standard deviations of these degree changes from
horizontal, with the overall stability index being for changes with respect to both AP and ML
indices (149). Velocity is in reference to the speed at which a COM moves through its pathway
deviations (52). Sway path (or postural sway) is defined as taking the anterior-posterior and
medial-lateral postural sway lengths to calculate the total postural sway, measured through
weight distribution and pressure shifts, often using a stationary force platform (51). It can also
be calculated through 3D motion capture methods where the COM movement is tracked in the
anterior-posterior and medial-lateral directions for total displacement (52).
Postural sway has been significantly reduced post-operatively compared to pre-operative
levels (51). The TKR patients still tend to have an increased sway path compared to the healthy
controls after their operation. Prior to any sort of training program, the TKR patients have
increased anterior-posterior and medial-lateral sway path when their eyes are opened (48). In
comparison to controls, these increased sway paths for the TKR patients have lasted through 6
months post-operation. The anterior-posterior sway path for the operated limb has been shown
to be increased compared to healthy controls at the 6 month mark. The increased anterior-
64
posterior sway path has also been shown to be associated with an increased equivalent area,
which is defined as the area of the COP path (139), calculated through a temporal sum of the
movement of the path in every direction (150). Interestingly though, when testing with their
eyes closed, the TKR population has small differences in sway path and area, but no difference
in AP and ML sway with no training compared to healthy controls (48).
Training programs specifically geared towards balance improvement have been shown to
reduce balance deficits in the TKR population (48, 149). As early as 6 weeks after a TKR
procedure, a balance training group showed an improved overall stability index compared to
their own pre-operative levels. Additionally, the same group showed an improvement in overall
stability index and anterior-posterior stability index in comparison to a control group that
received no balance training (149). Another training group showed improved balance on both
the operated and non-operated limbs in a single leg balance test for conditions with the eyes open
and closed (49). Differences have been shown between the two limbs of the TKR patients.
Small medial-lateral velocity of the COM (defined as the speed of the movement through the
displacement of the COM) increases have been shown during single leg stance on the operated
limb compared to the non-operated limb, but during a bilateral standing task, no differences were
reported between the two limbs (52), suggesting a slight increase in ability of the non-operated
limb.
During an examination of related balance variables, pre-operative postural sway was seen
as a strong predictor of single limb static balance post-operatively for TKR patients (51). In
general, improvements in balance have been associated with improvements in gait speed, timed
up and go tests, 30 second chair rise tests, and stair climb tests (49). The argument may be made
65
that people may be able to avoid single limb static balance situations, however, the connections
between these balance abilities and other functional abilities are evident.
Strength and balance are frequently connected since strength is necessary to maintain
balance. Increased knee extensor strength coupled with an increased gait speed has led to
increased anterior-posterior balance (measured through the range of the AP trajectory for the
COP), but increased knee extensor strength with a reduced gait speed led to a reduced anterior-
posterior balance (50). Peak torque did not predict single leg static balance performance (51).
Conversely to this point, another study has found that a failure to maintain single limb balance is
explained by older age, higher body mass index, and reduced quadriceps strength (52). These
unilateral differences do not appear to be related to TKR prosthetic design. A comparison of
cruciate retaining versus cruciate substituting designs showed no differences in single-limb
stance balance scores (143), suggesting that a retention of the posterior cruciate ligament has no
bearing on the balance performance improvements post-operatively.
While balance on a moveable platform may not be considered a “real life” situation, the
connections of balance to functional improvement have been well demonstrated. This suggests
an examination of balance as an important factor in return to functional improvement. A
rehabilitation program post-operatively should include a balance component given the
improvement seen in the functional tests with improved balance skills. Currently, as with
strength and biomechanics, there is a lack of research regarding TKR satisfaction levels and how
it relates to balance variables. An instability may cause individuals with TKRs to not feel
comfortable on their feet due to fear of falling, which may lead to an overall dissatisfaction. It is
important for investigators to examine balance and how it relates to dissatisfaction, as this is a
66
trainable physical skill. Training, as shown earlier, can improve balance skills in the TKR
population (48).
Creation, Examination, and Application of Survey Tools
Patient reported outcome measures (PROMs) are commonly utilized tools in the
treatment and rehabilitation process for people with knee osteoarthritis and total knee
replacements. PROMs are commonly utilized in conjunction with objective doctor analyses
regarding the injury or pathology because the patient’s perception of the outcome is considered
as important as the doctor’s objective measures. However, there are some issues with respect to
the PROMs being implemented. In knee arthroplasty alone, there are over 25 survey instruments
measuring patient outcomes being applied to the population (151). Based on the opinion of one
author, of those 25, only three have been extensively studied with respect to their validity in the
research field: the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC),
the Oxford Knee Score (OKS), and the Short Form 36 (SF-36) (151). The validity studies of
these tools encompasses the assessment of multiple psychometric properties such as content
validity, internal consistency, construct validity, criterion validity, agreement, reliability,
responsiveness, floor/ceiling effects, and interpretability. These criteria are assessed based on
examination of the creation of the tool, statistical analyses of the results, and application to the
population for which it was intended. Often, however, these measures are utilized to assess
outcomes in populations for which they were not designed. The SF-36, for example, was created
as a general health survey for any population while the WOMAC and OKS were created
specifically for an osteoarthritic population (152-154). These tools are all used to help assess
patient satisfaction with a medical procedure, recovery, or the rehabilitation of the injury. In
67
doing so, it becomes important to assess patient satisfaction and what that means in addition to
the survey tools being implemented to help do so.
Psychometric Properties
As surveys are developed, they are frequently tested for their abilities to perform well
under certain measurement conditions. They are evaluated on their “quality criteria” which
essentially determines their overall validity as a tool. Quality criteria includes assessments such
as content validity, internal consistency, criterion validity, construct validity, reproducibility,
responsiveness, and floor/ceiling effects (155). Each tool has multiple measurements of each
aspect of the quality criteria, allowing for an overall assessment of the tool based on the results.
Content validity is assessed by the relevance of items in a specific domain being
measured on a specific population to whom the tool is applied as well as the adequacy of
questions within those domains in being a reflection of the true purpose of the tool (156). The
presence of floor or ceiling effects and skewed data can reduce content validity. This may be an
insufficient measurement element for content validity as an imposition of limits is applied to the
constructs being assessed, such as the constructs within the tool (151). As a whole, the disease-
specific questionnaires tend to be less skewed than the general health questionnaires (157).
Unfortunately with a multitude of means for assessing any individual psychometric
characteristic, there will be multiple responses for whether a tool possesses a high value for any
given characteristic. The Forgotten Joint Score, for example, is said to have content validity for
the tool as a result of including the patients as part of their expert panel during the development
of the tool. Therefore because their opinions were included during the construction of the
content, the tool has content validity (158). This may seem like a simple way of assessing
68
content validity, but it is a published and accepted mechanism, regardless of how subjective or
imprecise such opinions may be.
Internal consistency is a measure of the degree of correlation or homogeneity of items in
a tool (151). There are two commonly used methods for examining the internal consistency of a
tool: factor analysis and Cronbach’s alpha. Factor analysis is a process of examining data in an
attempt to find patterns and dimensionality. Similar items within the tool are expected to provide
the same dimension (159). There are two different options: an exploratory factor analysis and a
confirmatory factor analysis. An exploratory factor analysis is performed when there are no
clear-cut ideas about the factor structure (no clear number of dimensions or what their
associations are). A confirmatory analysis is performed when prior hypotheses exist, whether
those hypotheses are based on theory or on previous analysis (160). The use of Cronbach’s
alpha statistic is more commonly reported in the literature in its use for evaluating internal
consistency. “Cronbach’s alpha estimates the degree of equivalence between responses to sets of
items tapping the same underlying concept. The higher the alpha, the higher is the average
correlation between responses to all possible combinations of items in the scale” (156).
Construct validity is a description of the relationship of the tool in question to other
measures which are attempting to assess the same underlying variables (also known as the
constructs). Theory-based hypotheses are created prior to correlation measurements to determine
the relationship of the measures (155). When a survey tool receives a positive rating for
construct validity, it is typically as a result of containing a priori hypotheses and a confirmation
of the majority of those hypotheses (151). A failure to provide a priori hypotheses often reduces
the construct validity rating (161, 162). The generation of predefined hypotheses is important as
it reduces the bias in the data. Without them, there is a tendency to look at the low correlation
69
values for explanations of the construct validity rather than coming to the conclusion that the tool
does not have an adequate level of construct validity (155). This seems to be more of an
individual study issue as pre-defined hypotheses can be created and be less of an issue with the
actual tool.
Criterion validity is the examination of how scores on a given instrument relate to a “gold
standard” (163). This seems to be a rarely examined psychometric within the TKR population as
there does not seem to be a general consensus on what the “gold standard” is for the field with
respect to surveys utilized to assess pain, function, expectations, and other factors. This has both
good and bad elements to it as it allows for the use of multiple surveys to examine different
populations and pathologies, but it makes comparisons difficult.
Reproducibility is the extent in which multiple measurements of a patient can produce
similar results when no real change has occurred. Variations in the subject or rater could cause
changes in the measurements when real change is not present over time, thus contributing to
background noise and subsequently affecting the reproducibility of the tool (164). There are two
components of reproducibility: agreement and reliability. Agreement references how close
scores are on repeated measurements and estimates the absolute measurement error. Agreement
is based on the calculation of the standard error of measurement (SEM) and the relationship of
the smallest detectable change (SDC) and the minimal important change (MIC, also
interchangeable with minimal clinically important difference; MCID). The SEM can be used to
derive the SDC. The SDC represents the smallest change in score that can be interpreted as real
change beyond measurement error (155). Relating the SDC to the MIC for comparing
agreement has been performed but is a newer construct and has not been used in most studies
(155). When examining the SDC, the threshold for detecting these “just noticeable” differences
70
is often 0.5 SD (165), which provides a quantifiable difference. When designing studies, it is
important to consider that assumed effects below the MCID may be detectable but are essentially
clinically meaningless (166). Reliability parameters relate the measurement error to the
variability between subjects and thus rely on the heterogeneity of the sample, while agreement
parameters, being concerned with measurement errors, reflect the characteristics of the
instrument itself (167). Some methods of assessing reliability are considered inappropriate, such
as the use of Pearson’s correlation coefficient, rho, and paired t-tests (151). The use of these
“inappropriate” methods is believed to cause misleading interpretations (168).
In comparing agreement and reliability over different population samples, agreement
parameters will be more stable than reliability parameters. Reliability is highly dependent on the
variation that exists within a population sample and it is essentially only generalizable to samples
which possess a similar variation. Reliability is more of a characteristic of how a tool performs
in a certain sample population and is a particularly vulnerable concept in self-rating measures.
Agreement is more of an element of the tool itself, and not the population. When the instrument
is to be used for evaluation purposes, as it will with most medical based research such as TKR
assessments, agreement is the preferable mechanism for evaluating reproducibility (167).
There have been multiple definitions and multiple methods of measuring responsiveness of a tool
utilized, which unfortunately shows a lack of consistency for use with standardized terminology
and approach (169). The goal of measuring responsiveness is to test the tool for its ability to
respond to change that has occurred in the patient, but it is important to note that a general
change can be seen as different from “clinically important change”, in the sense that statistically
significant change does not necessarily indicate any clinical relevance (151). For example, large
sample sizes can result in small numerical differences which are seen as statistically significant
71
(170), but this does not mean that the changes have any clinical meaning. The use of the MCIC
to gauge clinically important change is commonly utilized as it quantifies changes in the tool
which patients perceive as beneficial (171). As previously mentioned, there are multiple
methods of measuring responsiveness (effect size, standardized response mean, Guyatt’s
statistic, Paired-sample t-test, and relative efficiency) (172), which makes the direct comparisons
tough as each statistic measures something different despite trying to explain the same thing,
thus only allowing for general comparisons of responsiveness.
There are two broad approaches to defining clinically meaningful change: distribution-
based and anchor-based. There are conceptual differences between the two methods.
Distribution-based is centered on statistical criteria, while anchor-based examines the clinical
relevance of measures. The examination of distribution-based approaches shows that the study
sample characteristics and the use of standardized response means, effect sizes, and t-statistics
define change. Conversely, for anchor-based approaches, the changes examined and linked to an
external, yet relevant, clinical anchor. These external anchors can be something like a global
rating or disease-related outcome and are used to assess whether individual changes possess any
clinical significance (173).
A floor effect is when a patient scores the lowest possible score on the tool, suggesting
they may actually fall outside of the scoring range. A ceiling effect is the exact opposite where a
patient scores the highest possible score, also suggesting they fall outside the upper limit of the
scoring range. The presence of floor and/or ceiling effects can affect multiple aspects of a tool’s
validity, including content validity, reliability, and the ability to detect change. If either of these
elements are present, it may mean that the tool is missing the extreme items on either end of the
scoring spectrum, thereby suggesting the content validity is limited. This will then decrease the
72
ability of the tool to delineate patients who have the highest scores (ceiling) from the ones who
have the lowest scores (floor) and its ability to detect further deterioration or improvement for
the patient (155). Essentially, this means that the tool fails to recognize differences between
patients who fall outside the range of measurement of the tool, either above or below it (163).
Disease-specific surveys score better for floor and ceiling effects than general-health surveys
(151, 157, 174). This is a plausible outcome as the application to a TKR population should
rarely produce ceiling effects given that a return to full function after a TKR procedure is
unlikely. Floor effects are more likely to be experienced as a severe disability is what leads to
the TKR procedure so depending on the time of application of the tool, floor effects are possible.
The closer to the procedure the tool is administered, the more likely a floor effect.
Interpretability is the ability to change quantitative scores into some sort of qualitative
meaning (163). Surveys can receive positive ratings for interpretability for multiple reasons,
such as the presentation of a mean and standard deviation, inclusion of multiple relevant
subgroups, and the relationship of scores to clinically relevant conditions (151). Interpretability
is less a comparison of the different tools to each other and more an aspect of how the
information and statistics are presented from the data collected. Each tool mentioned has the
ability to have good interpretability with the right amount of data being presented and in the
correct manner.
Development of Surveys Used on TKR Populations
WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index). The
WOMAC is one of the more widely used tools for the TKR population to measure pain, stiffness,
and physical function (www.womac.org). It has been researched and validated for paper,
telephone, computer mouse, and touchpad administration and translated into over 60 languages
73
(175). This tool was developed with the knee and hip osteoarthritic populations in mind. It
examines pain (5 items), stiffness (2 items), and physical function (17 items) using an ordinal
scale range from 0-4, with lower scores indicating reduced physical function and increased
symptoms. This allows for both subscale (pain, stiffness, and physical function) and global
(total) scoring. The subscales are summed, with maximum scores of 20 (pain), 8 (stiffness), and
68 (physical functioning). The global score is the sum of all subscale scores, with a higher
scores indicating increased function and reduced symptoms related to the knee joint (176).
There is also an option to use a visual analogue scale rather than the ordinal scale for responding
to the questions.
The WOMAC was developed by a team of four rheumatologists and two clinical
epidemiologists. Initially the team developed open-ended questions to examine the
characteristics and clinical importance of pain, stiffness, and physical, social, and emotional
dysfunction. Closed-ended questions from existing surveys (Functional Status Index, Pilot
Geriatric Arthritis Project) were added in an effort to complete each dimension by finding any
sources of discomfort or disability (152). Using a 0-4 scale, patients were asked to report on the
presence or absence of any discomfort or disability, the frequency of the discomfort/disability,
and the importance of the discomfort/disability to them. All questions were asked with respect to
what was only specifically linked to the osteoarthritis. Items that were seen as sex-specific in the
early 1980s were avoided (such as ironing) and were rephrased to more generic terminology
(such as light domestic duties). One hundred patients were interviewed following the final
creation of the tool (152), which ultimately left out the social and emotional components. Social
dysfunction had mean importance scores of 2.2-2.7, similar to other scores of discomfort and
disability. However, it was eliminated due to it having low occurrence rates of 3-27%. While,
74
emotional dysfunction also had similar levels of mean importance (2.1-2.6), a low prevalence (9-
56%) was seen. The low occurrence of social and emotional dysfunction led to their elimination
from the tool.
Following the final development, the tool was tested for reliability, responsiveness, and
validity for TKR and total hip arthroplasty (THA) patients as well as osteoarthritis (OA) patients
receiving non-steroidal anti-inflammatory drug (NSAIDs). This allowed for application to both
OA and joint replacement patients (177, 178) which is an asset to the tool as TKR is traditionally
the treatment option for end-stage OA. It has been tested with other therapeutic interventions
(acupuncture) and for patient groups other than knee and hip patients, but with much less
frequency (176).
The WOMAC was intended for an osteoarthritic population, which is linked with TKR,
but has been applied in wide usage to OA. This may contribute to reduced content validity. The
WOMAC has scored well on content validity, but has been reported as lower than other disease-
specific surveys (157, 179, 180). A subsequent examination of floor and ceiling effects and
skewness of data distribution may help explain the WOMAC scoring as this is also a common
assessment for content validity (157). The wide application of the WOMAC has led to the
presence of floor and ceiling effects, leading to reduced content validity surveys (157, 179, 180).
The WOMAC has been tested for internal consistency with both factor analysis and with
the use of Cronbach’s alpha. In a Singaporean TKR population, the Chinese and English
WOMAC versions resulted in five and seven factors, respectively (181), which does not support
the factor structure because there are not that many subscales in the original version. Another
study on the WOMAC showed that the pain and function subscales were not different from each
other (182), indicating reduced internal consistency. As a result of this, some research has
75
chosen to investigate the individual subscales independently rather than performing a single
analysis (183). The pain subscale has been further shown to lack unidimensionality (184), also
implying poor internal consistency. Using Cronbach’s alpha, the WOMAC had an alpha score of
0.93. demonstrating good internal consistency (157).
The WOMAC has received mixed ratings for construct validity, with the reduced
construct validity coming as a result of failure to provide a priori hypotheses (151, 161, 162).
Effect sizes are sometimes utilized to assess construct validity. The WOMAC pain subscale has
shown a high effect size (0.95) while stiffness and physical function had moderate effect sizes of
0.78 and 0.76, respectively (162).
Researchers have shown positive and indeterminate ratings for reproducibility for the
WOMAC. When reporting reproducibility based on random effects intraclass correlation
coefficients (ICC), the WOMAC had ICCs above 0.9 (34, 157). Test-retest reliability on the
WOMAC also yielded positive results, suggesting good reproducibility (162). There has been
research done in which the clinically important difference (CID has been examined on the
WOMAC and been correlated with a subjective assessment response of “a good deal better” with
respect to the post-surgical improvement (185). In rehabilitation interventions, the MCID for the
WOMAC was reported as 12% of the baseline score or 6% of the maximum score for detecting
differences (166). Effects of this size lead to smaller required sample sizes (<300). The larger
the effect, the smaller the necessary sample size.
An examination of responsiveness has shown that the WOMAC pain and physical
function scales were both responsive to clinical change for knee patients (186, 187), “Relative
efficiency” has been compared between WOMAC and the SF-36 (a general-health survey), with
the WOMAC scoring better than the SF-36 post-surgery for a TKR population, although the
76
score decreased with time, meaning that initially the WOMAC was more responsive but the
values moved closer to each other over time (172, 188). This makes sense as time away from
surgery should hopefully see a return to normal health, and thus similar responsive values as the
magnitude of change decreases. This seems to be a common trend in which the disease-specific
tools outperform the general health scales, as they should, given their narrower focus. Disease-
specific tools have increased responsiveness (162, 172, 188) as a whole.
Researchers have shown mixed results for floor and ceiling effects for the WOMAC.
Ceiling effects have been reported between 16.7-46.7% and floor effects between 0.4-0.8%
(158). In a different study, the WOMAC showed neither ceiling nor floor effects (162), which is
contradictory to other research, indicating that it may be a population-specific issue. The
disease-specific surveys tend to score better for floor and ceiling effects than do the general-
population surveys.
In studies using the WOMAC, means and standard deviations are commonplace for
reporting (162, 189) as are reporting at multiple measurement time periods, which gives clinical
relevance as it tracks changes over time (34, 188). This provides high levels of interpretability as
these quantitative scores can be interpreted with qualitative meaning (163).
FJS (Forgotten Joint Score). The Forgotten Joint Score (FJS) tool was created using
similar methods to both the WOMAC and OKS (158). An initial 20 question pool was created to
measure joint awareness based on literature research describing the important elements of a
TKR/THA procedure and the analysis of a team of expert opinions from clinicians, a
methodologist, and a statistician. Patient opinions were then consulted based on the importance
they placed upon specific activities of daily living (ADLs). After pilot testing the 20 questions
and breaking them down for individual analysis, 6 questions were initially eliminated due to a
77
high percentage of missing items. The remaining 14 questions had high internal consistency
(Cronbach’s alpha of 0.96). No item significantly reduced the alpha level, so nothing else was
deleted. Twelve questions were selected to complete the final tool (Appendix A), with two
questions being combined with other questions to place different sporting activities together
(158). The FJS uses a five point Likert scale for the scoring of the 12 questions, with each
question assessing awareness of the replaced joint during specific activities (sitting, walking,
bathing, etc.). The five points represent a response of “never, almost never, seldom, sometimes,
mostly” in relation to the awareness question (158). The goal of the tool is to assess whether or
not the patient is able to forget the joint is replaced during their everyday lives (163), hence the
examination of the awareness of the joint during ADLs. The ability to forget the joint exists had
been deemed, by the creators of the tool, the ultimate measure of satisfaction (158), although this
does fail to take into account pain, social, and mental constructs which may inhibit satisfaction
levels.
In assessing the psychometric properties of the FJS, there is much less research
concerning this survey compared to others as it is relatively new. The FJS had a high alpha
(0.95) upon its creation and it was reported that none of the 12 items significantly lowered the
internal consistency of the tool (158). A subsequent translation of the FJS to Japanese yielded an
alpha of 0.97 (190), thus further supporting its internal consistency. Correlations between the
FJS and the WOMAC were reported by its creators as -0.69 (opposing scoring methods result in
negative correlations) or greater for all subscales and total scoring for the WOMAC, with the
total score having an r equal to -0.78 (158). However, when translated to Japanese, correlations
to the WOMAC ranged between 0.289 (pain subscale) and 0.522 (total score), indicating a
78
reduced internal consistency across languages. However, the Cronbach’s alpha remained high at
0.97, indicating an acceptable level of reliability (190)..
When creating the FJS, the creators sought to eliminate ceiling effects by expanding their
responsive values beyond a “good” response to an “excellent” response (158). They still
reported a ceiling effect of 9.2% which they reported to be lower than the WOMAC. Although
floor effects of 3.3% were higher than the WOMAC, which is expected results since the FJS or
for replacement patients while the WOMAC can be for anyone along the OA spectrum (158).
Other studies have reported the ceiling effect for the FJS as high as 40%, more specifically for
the first five questions of the tool (which are the easier ADLs, therefore the presence of ceiling
effects is logical), and a floor effect for 16% of the patients, with significant differences between
types of knee replacement procedures (191). The authors did speculate, however, that the
presence of a ceiling effect post-surgery may mean reduced chronic pain levels, which is a
positive attribute (191).
OKS (Oxford Knee Score). The OKS is a 12-item questionnaire developed specifically
for patients undergoing a TKR to assess pain and functionality (34). During its creation, 20
patients were interviewed about their experiences and problems with their knees during the time
of surgery. From these interviews, 20 questions were developed. These questions were then
given to 20 additional patients and they were asked to add their comments regarding the
questions and problems they experience which they felt were not addressed by the survey.
Subsequent adjustments were made to the questions in the survey. The adjusted questions were
then given to another set of patients. The same process was repeated for a total of three rounds,
leading to the final development of 12 questions to complete the survey examining pain and
function. There is a 1-5 scoring scale relating to difficulty and/or pain levels with certain
79
activities. Each number has its own description of a corresponding answer to the question on the
survey. For example, a question about pain may have anchors of “none (1)” to “severe (5)”
while another question has anchors of “Not at all (1)” and “Totally (5)” (34). This survey is
often misinterpreted as the score scale is inversed from many other tools, in that a lower score
represents a more positive outcome. There is a total possible score of 48. Scoring starts with a
total score of 60 and points are subtracted from that based on patient responses to questions. A
score of 0-19 may indicate severe OA, 20-29 is moderate to severe OA, 30-39 is mild to
moderate OA, and 40-48 indicates satisfactory joint function (163). The survey has many
benefits in that it is short, practical, and reliable, leading to it being used more frequently in
assessing the TKR population (18, 192-195).
The OKS had mixed results with respect to content validity, with some researchers giving
it high, positive ratings (34, 196) and others gave the OKS lower content validity scores (157,
174, 189). In further assessing content validity, when scoring it based on floor/ceiling effects
and skewness, the OKS had improved floor and ceiling effects but a lower skewness rating,
thereby lowering its content validity rating (157).
In order to measure internal consistency, factor analysis on the OKS has been performed
on three different factors for two different language versions (English and Chinese). The English
version had pain on one factor, physical function on another, and a combination of the two on the
third. The Chinese version had a combination of pain and physical function onto the first factor,
limping onto another, and kneeling and night knee pain on a third (196). This illustrates that
there can be differences with internal consistency which translate to beyond just the number of
factors and how they align, making the creation of those factors an important construct to
consider when evaluating them. The factor analysis performed on the OKS was an exploratory
80
factor analysis (196), which seems to be the appropriate method as the OKS does not have
clearly defined subscales, whereas the WOMAC does and should therefore be used under a
confirmatory analysis.
An examination of internal consistency with Cronbach’s alpha has shown the OKS had
an alpha score of 0.93 (157). Two other studies have shown the OKS value to be greater than 0.9
(161, 174). An examination of the time of application for the OKS revealed scores of 0.87 prior
to TKR operation and 0.93 six months after the operation (34). When divided into subscales of
function and pain, the OKS had alpha levels of 0.819 and 0.874, respectively (161) (It should be
noted that the OKS does not technically have subscales, as mentioned earlier, however in
examining the questions, they are fairly easily placed in a function or pain category based on the
wording).
The OKS has received positive ratings for construct validity, when providing a priori
hypotheses, however when failing to do so, the construct validity scores have declined (161,
162). The OKS also displayed moderate correlations with the SF-36 physical function subscale.
The OKS function and pain subscales (again, not designed with specific subscales, but have been
devised by researchers after the creation of the OKS) had Pearson’s correlation coefficients of -
0.69 and -0.72, respectively (the correlation values are negative not due to an inverse relationship
but due to the inverse scoring methods between the two tools. The physical component summary
score had values of -0.73 and -0.76 with the same two OKS measures (161). The correlation
values did drop as low as 0.19 between the OKS and the mental health component summary
score of the SF-36 (34), which is to be expected and suggests good divergent construct validity
(174).
81
Researchers have shown mixed results for reproducibility of the OKS. The OKS
received ICCs above 0.9 (34, 157). The negative results are attributed to the use of inappropriate
methodologies, such as the use of Pearson’s correlation coefficient, rho, and paired t-tests (151).
It is believed that these measurements cause misleading interpretations (168). Additional tests
examining responsiveness of the OKS showed effect sizes in excess of 2.1, suggesting good
responsiveness of the survey. This score was higher than other surveys such as the SF-36 (34).
Floor and ceiling effects have been found for the OKS. The OKS had a 6.8% floor effect
and a 0.1% ceiling effect (174). The OKS scored better for the presence of floor and ceiling
effects compared to the SF-36 (151) and the WOMAC (157). This indicates the increased ability
of the OKS to better measure the TKR and OA populations without the population failing to
accurately score due to being too high or too low.
High interpretability scores have been given to the OKS. Similar to the WOMAC, means
and standard deviations are commonly reported with the OKS (162, 189) as are reporting at
multiple measurement time periods, giving clinical relevance to the information as changes are
tracked over time (34, 188). This allows the change of quantifiable information into clinically
useful qualitative information, which contributes to the increased interpretability (163).
KOOS (The Knee Injury and Osteoarthritis Outcome Score). The Knee Injury and
Osteoarthritis Outcome Score (KOOS (197)) is more comprehensive than the WOMAC, OKS,
and FJS in that it assesses more domains. It was designed to assess five different outcome
measures: pain, symptoms, ADLs, sport/recreation function, and knee-related quality of life.
Like the FJS, a literature search was performed in an attempt to identify areas of importance for
patient-relevant outcomes, which led to subscale areas of symptoms, functional status, and
satisfaction. An expert panel was then created consisting of patients, surgeons, and physical
82
therapists. They were asked to identify both short and long term symptoms and functional
disabilities from meniscus and anterior cruciate ligament injuries, although they desired to have
it apply to early OA patients as well, with the reasoning that meniscus and ACL injuries can lead
to OA. From this consultation with the expert panel, seven identifiable factors emerged: pain,
early disease specific symptoms, late disease specific symptoms, function, quality of life, activity
level, and satisfaction. Pilot testing led to the scale being narrowed down to five outcomes, with
satisfaction and activity level being left off of the final survey due to the failure to agree to
wording which would be relevant to all knee-related situations. For example, the authors could
not agree on wording for a question which could be posed to an ACL patient as well as a TKR
patient. A total of 42 items were included with a scoring system from 0-4, with higher scores
meaning less problems with the knee. Each subscale was scored independently: pain (9 items),
symptoms (7 items), ADLs (17 items), sport/recreation function (5 items), and knee-related
quality of life (4 items) with each subscale having its own anchors for the Likert-scoring system.
Scores were then transformed for each subscale into a 0-100 range. Unlike other surveys with
global scoring, a total score was not considered because each individual item was deemed
important enough for separate analysis and interpretation (197). In reality a global score could
easily be calculated through simple summation of the individual subscales, however, this was not
the original intention of the tool and should theoretically be done with caution.
In assessing the psychometric properties of the KOOS, content validity has been reported
as high through comparison with the WOMAC (198). The creators of the KOOS implemented a
method for calculating the WOMAC scores through the KOOS subscales as a means of insuring
content validity. The KOOS physical function component had an alpha of 0.89 and a
corresponding correlation with the WOMAC physical function of 0.9, suggesting good internal
83
consistency (199). When applied to five different levels of OA (mild, moderate, severe, TKR,
TKR-revision), the KOOS had alpha levels for all subscales above 0.7 with only one exception
(symptoms subscale for the severe OA group; 0.56) (200). This suggests that different subsets of
a population are more sensitive to specific subsets of the tool, indicating that care should be
taken with the interpretation of the results. It has been stated that it is tough to make a
generalization on internal consistency as data on dimensionality and factor structure is limited
for most of the tools (151).
The KOOS had a high correlation (0.9) with the WOMAC in the physical function
subscale (199), indicating similar constructs and good construct validity, which is to be expected
given that both are disease-specific tools and investigating similar things. When comparing the
KOOS with the SF-36, the highest correlations between the two were on comparing physical
function to ADLs, obtaining a value of 0.57, which can be interpreted as a moderate correlation
(197). This is likely in the sense that physical function and ADLs are slightly different and may
not be exactly the same. Additionally, the KOOS and SF-36 are different surveys measuring
different elements. Higher correlations are seen with the physical components of the SF-36 than
the mental components compared to the KOOS (197), which should happen given that the
KOOS examines physical components, and not mental components.
The criterion validity is less commonly researched. One study has used the WOMAC as
the gold standard in comparison to the KOOS to assess criterion validity (199). While the
WOMAC is the most commonly utilized tool, it is seldom referred to as the gold standard. The
KOOS had standardized response means ranging from 1.4 to 1.7, indicating good responsiveness
in comparison to the WOMAC (199). Furthermore in comparing the KOOS to the WOMAC and
SF-36 for responsiveness, the effect sizes were all high for the KOOS subscales (0.8-1.64),
84
which was higher than any of the WOMAC or S-36 scores, except for the knee-related quality of
life scale (it was the most responsive subscale at 1.64) as this was deemed to have no replicated
measure in either the WOMAC or SF-36 (197). Initial testing of the KOOS found high random
effects intraclass correlation coefficients (ICC): 0.85 for pain, 0.93 for symptoms, 0.75 for ADL,
0.81 for sport/recreation function, and 0.86 for knee related quality of life (197), which indicates
high levels of reproducibility for all the subscales.
Floor and ceiling effects have both been reported for the KOOS. When examined in
multiple stages of OA, researchers showed the KOOS had a ceiling effect for the mild OA group
for pain, symptoms, and ADLs and for sport/recreation for the severe OA group and also had
floor effects for sport/recreation and quality of life in the severe OA and TKR revision groups
(200). This suggests that the interpretation of the specific group it is applied to is important as
well as the subscales within those specific groups. A ceiling effect may be achieved with
disease-specific tools but at the same time, pain, for example, may continue to occur elsewhere
as a result of an existing comorbidity (201). When the presence of ceiling or floor effects exist,
additional examinations into other areas or extended scoring scales may be necessary.
Short Form 36 and Short Form 12. The WOMAC, OKS, FJS, and KOOS were all
created with either a specific disease and/or a specific joint in mind. However, not all tools
utilized on the TKR population were developed in this manner. The Short Form 36 (SF-36) and
the Short Form 12 (SF-12) are general health surveys which survey a wider array of information
than the disease-specific surveys, but are often applied to the TKR population in conjunction
with other surveys (17, 202, 203). These general health surveys are seen as more encompassing
of an overall picture of health. The SF-36 measures three aspects of health: functional ability,
well-being, and overall health. Within these three aspects are 35 questions divided into 8
85
subcategories: physical function (10 items), bodily pain (2 items), role-physical (4 items),
general health (5 items), vitality (4 items), role-emotional (3 items), social functioning (2 items),
and mental health (5 items) (204). An additional question regarding overall health status as
compared to one year prior is also included, to give the total of 36 questions (154). Responses
are scored on a 5 point Likert scale. Scores can be converted to a total score between 0-100
(worst to best) to aid in interpretability. Subscale scores can be provided by summing the
responses from the questions in each subscale. A physical component summary score and a
mental component summary score are provided when combining subscales. The physical
component consists of the physical function, role physical, bodily pain, and general health while
the mental component consists of mental health, role emotional, social function, and vitality
(175). The SF-36 was created over a seven year period with the intention of improving and
updating the 18- and 20-item Medical Outcome Survey (MOS) short form. The goal was to
create a more efficient scale for measuring general health (154). It was designed based on
previous surveys examining patient limitations in physical, social, and role functioning, general
mental health, and general health perceptions. These previous surveys were deemed too long
and thus required a shorter, yet still comprehensive, version in order to capture the total picture
of health (154). Each independent subcategory had its own construction process, ranging from
exact replicas of previous surveys (physical function, mental health) to completely new
constructs (vitality) and everything in between (154).
The SF-36 is a generic health survey which allows for comparisons between different
patient groups with the same condition as well as different conditions. The problem though is
that is does not examine the intimate details of specific injuries/diseases, leading to less insight
about the specific nature of problems with a patient for a given issue. When creating this survey,
86
breadth (issues with the comprehensiveness of the tool) and depth (issues with precision in
measuring a specific concept) were seen as potential problems. Breadth of the tool was
addressed by examining and including the most frequently studied functional status aspects and
well-being concepts in previously described and widely accepted definitions of health status.
Depth of the tool was addressed by taking the subset of items which best replicated the full
length tools. These tools already had proven validity, and thus could be applied to the shorter
version (154).
While the SF-36 was created to make a shorter version of a measurement tool, it was still
longer than some other tools (WOMAC), leading to the creation of an even shorter version which
would become the SF-12. During the creation of the SF-12, the creators had three goals in mind:
1) create a form that could be scored to explain at least 90% of the variance in the physical and
mental health measures of the SF-36, 2) create a form that could reproduce the average summary
scores and the eight subscale scores with a high degree of comparability, and 3) create a form
which could be self-administered on two pages or less or administered by a tester in less than
two minutes on average. The creators used data from the National Survey of Functional Health
Status to select and score 12 items from the SF-36 and the Medical Outcomes Survey and to
cross-validate population-based scoring algorithms for both the summary measures and the
subscale measures. It took ten items to produce physical and mental component summary scores
of the SF-36 with R2 values of 0.9 or higher (physical=0.911, mental=0.918). An additional two
items were selected, thus allowing the 12 chosen items to represent all eight subscales (205).
As a whole, the OKS and KOOS have shown higher positive ratings for content validity
compared to the SF-36 (157, 179, 206) and SF-12 (157, 179), as indicated by increased presence
of floor and ceiling effects. The lower ratings for the SF-36 and SF-12 are likely respective of
87
the populations with which they were applied to. Since they are general health surveys whereas
the OKS and KOOS are disease-specific, application to a TKR population should warrant higher
scores for the OKS and KOOS. An assumption of content validity for the SF-36 and SF-12 may
not be appropriate when applying them to a TKR population as they were not originally intended
for use in that population (151). The target population is specific for the OKS and KOOS while
the SF-36 and SF-12 can be applied to anyone.
On the SF-36, a variety of methods, presentations, and interpretation of factor results
have been shown (160). Most of the studies that examined the factor analysis of the SF-36 used
exploratory analysis instead of the confirmatory, which would have been a more appropriate
choice (160). However, when examining the SF-36 and SF-12, it is important to note that they
were constructed with heterogeneity in mind, not homogeneity, based on the generalization of
the survey and not the disease-specific element (205). This may indicate overall poor scores of
internal consistency given the lack of homogeneity. In using Cronbach’s alpha to assess internal
consistency, the SF-36 and SF-12 have shown good positive results. The SF-36 had an alpha of
0.85 and the SF-12 had 0.88 (157). All subscales for the SF-36 had alpha values of 0.75 or
greater while the SF-12 mental and physical component summaries had alphas of 0.62 in one
study (157). The SF-36 sometimes displayed lower values for Cronbach’s alpha, but rarely
below 0.7 (162). The individual components however had a wide range of values, with a 0.651
for the vitality subscale and a 0.996 for the role physical subscale. The mental health component
summary score was reported as 0.511 while the physical health component summary score was
0.709 (161), further helping to illustrate the diversity within the tool.
Like the WOMAC, OKS, and KOOS, the SF-36 has received positive ratings for
construct validity (151). Lower ratings have been shown for the SF-36 when the researchers
88
have failed to provide a priori hypotheses (161, 162). In using effect sizes to assess construct
validity, values for all the subscales in the SF-36 were all 0.75 (physical function) or lower
(162). Generic measures better assess comorbidities than the disease-specific tools do (201).
This may be more beneficial during an aging process as these comorbidities are more likely to
occur, however it does not help with specific assessment of a specific issue, as the disease-
specific tools would. Examination of the comorbidities showed that the WOMAC did not
produce any significant differences in effect sizes but the SF-36 did for rheumatology patients,
but for TKR patients for both the WOMAC and the SF-36 in some of the subscales, but not all
(162).
The SF-36 and SF-12 have had both positive and negative ratings for reproducibility.
The WOMAC and OKS had ICCs above 0.9 while the SF-36 had a 0.75 (34, 157). As with any
statistic, the acceptability level depends on the interpreter, which means that despite the lower
rating for the SF-36, it can still be deemed a higher correlation value and thereby acceptable for
reliability standards. In a test-retest reliability for the SF-12, values of 0.89 for the physical
component and 0.76 for the mental component were reported (205). The same testing on the and
SF-36 also yielded positive results (162), suggesting good reproducibility for both the SF-36 and
Sf-12. In rehabilitation interventions, the MCID for the SF-36 was reported as 12% of the
baseline score or 6% of the maximum score for detecting differences (166). Each subscale of the
SF-36 is different, as the general health scale had the lowest MCID values for both individual
and group differences, being almost one-third the value of the other subscales (207).
The SF-36 has decreased responsiveness compared to the WOMAC (188). The SF-36
had large effect sizes for physical aspects of the scale (1.3-2.1), suggesting good responsiveness,
but only small to moderate effect sizes for mental (0.3-0.5) and social (0.4-0.6) aspects for a
89
TKR group (207). Additionally, the SF-36 showed better responsiveness at a group level
compared to an individual level for knee patients, and therefore may not be the best tool for
assessing change on an individual level (207). Some of the subscales of the generic measures
can have more responsive attributes, such as physical functioning, physical component summary
score, and bodily pain in the SF-36 exhibiting large effect sizes, thereby suggesting
responsiveness (172). These subscales are likely to be more responsive given the issues with
which they are being applied to (TKR) and that the physical aspects are the dominant area of
interest in this population.
During the creation of the SF-36, the presence of ceiling and floor effects were seen as
potential problems. In order to alleviate this, since it is a general health survey, the creators
suggested adding possible supplementary questions to the tool based on the population it is being
applied to (154). For example, if the tool is being applied to a severely diseased population, it
would be beneficial to add questions which better represent the extreme low end of the scoring
scale to control for floor effects, or for a very healthy population, add questions on the top end to
control for ceiling effects. The SF-36 showed ceiling effects for two of its dimensions: social
functioning and role limitations due to emotional problems, but no reported floor effects (162).
In a different study, the SF-36 showed a mean floor effect of 17.07%, with subscale values
ranging between 0.79% and 36.02%. while the SF-12 showed no mean floor effect, with the
subscales having 0.02% (157). In that same study, a mean ceiling effect of 12.52% was evident
for the SF-36, with subscale values between 0.59% and 49.50%, and the SF-12 had a -0.09%
ceiling effect. The ceiling effects in the SF-36 can be problematic as it does not allow for any
improvement (207), so depending on time of application of the tool, it may have limited ability
to detect improvement. To further this point, the physical function subscale had no ceiling
90
effects at baseline or follow up (207), which says something different. This may suggest an
inability to return to perfect physical form.
In the Sf-12 and SF-36, means and standard deviations are commonplace for reporting
(162, 189) as are reporting at multiple measurement time periods, which gives clinical relevance
as it tracks changes over time (34, 188). This suggests high levels of interpretability for both
survey tools, and thereby allowing qualitative interpretation of the information being presented.
The previously described surveys tend to be the most utilized surveys based upon
information collected for review articles on TKR populations (151). As a whole, they provide
either a disease-specific or a general health investigation of the patient. There are additional
surveys which have other items examined, such as ROM, other more specific activities
(triathlons, cross country skiing, etc.) and their difficulties, or a combination of activities in
conjunction with the self-reported patient measurements. For example, the Knee Function Score
examines specific physical therapy activities (timed walks, stride lengths, hop tests, kneeling
tests, etc.) and gives a score based on the patient’s performance of the activity while
simultaneously investigating the patient’s pain on an ordinal scale (208), however, it is much less
in depth with respect to any factor outside of the pain associated with the task (i.e. the difficulty
of a task). Theoretically, a task could be difficult without being painful so this tool lacks a
complete picture of the situation the TKR patient is experiencing. Other tools such as the New
Knee Society Scoring System looks at many more activities which require much higher levels of
physical functioning, such as jogging, weight lifting, racquet sports, etc. (209). Realistically, it
may take a combination of tools in order to successfully evaluate a patient as no single tool
seems to be all encompassing. With this in mind, as is evident in the decreasing size of the tools
91
being utilized, time is of great importance to everyone involved and the shortest, yet still
comprehensive, survey seems to be the desired entity.
Application of Surveys to TKR Population
A lot of good work has been done in the TKR population through the use of various
measurement tools. Physical function has been largely examined using the WOMAC, OKS, FJS,
KOOS, and many other tools designed specifically to address knee issues, while the SF-36 and
SF-12 have been used on the same population but examining a more general health assessment.
The SF-36 and SF-12 have mental health components to them, but they are broad in nature and
lack intimate depth in examining the psychological constructs. As has been shown, TKR
satisfaction extends beyond the physical into the psychological constructs. When applying
measurement tools to the TKR population, the WOMAC, OKS, FJS, or KOOS are important for
use as it allows comparison with other studies on the TKR population, but in attempting to
explain dissatisfaction, research going forward needs to include some psychological
measurement tool. Physical tools alone may not have the information needed to sufficiently
explain the dissatisfaction within a certain sample population. The addition of a psychological
assessment tool may help to further that explanation and provide a better understanding of why
patients experience satisfaction or dissatisfaction with TKR procedures. The survey tools may
be combined with additional physical testing which is not patient-subjective, such as
biomechanical assessment of walking and stair climbing. Biomechanical assessment has not
been done on satisfied and dissatisfied TKR populations. The incorporation of functional testing
commonly used in physical therapy may also explain the dissatisfaction in TKR patients. While
this creates a large study with the presence of physical surveys, psychological surveys,
biomechanical analysis, and functional testing, dissatisfaction is an important construct to
92
examine as the decision to undergo a TKR is not a small decision. It is a major surgery with a
long recovery. Patients should be able to expect satisfaction with the process if they are
investing this substantial amount of time and money into the procedure.
The research on the dissatisfaction has provided a substantial amount of information,
however there are some areas which need further examination. First, there is minimal
biomechanical data on the dissatisfied population. Kinematics and kinetics play a large part on
the evaluation of TKR patients, yet there has been relatively little information provided on the
kinematics and kinetics of the dissatisfied population and how they compare to satisfied TKR
patients and healthy controls. Additionally, it may be of interest to note how biomechanical
variables relate to strength, balance, and functional tasks for these same populations. This
enhancement of information available on the TKR population may help to dictate treatment of
dissatisfied patients. By identifying physical factors which may contribute to dissatisfaction,
researchers may be able to provide information for rehabilitating TKR patients towards
satisfaction with their joint replacement.
Patient Satisfaction
Meaning of Satisfaction
Satisfaction is a complex construct which could be argued as a psychological construct
rooted in physical properties, especially in pertaining to the TKR population. The vast majority
of the studies examining satisfaction in the TKR population have done so with a simple question
used to divide the population: “How satisfied are you with your knee replacement?”. The
patients are then provided with a 5 point scale which corresponds to options of very satisfied (5),
somewhat satisfied (4), neutral (3), somewhat dissatisfied (2), and very dissatisfied (1), possibly
with some slight wording deviations but still arriving at the ability to be able to distinguish
93
between satisfied and dissatisfied patients (9, 11, 17, 18, 111, 192, 194, 203, 210). A visual
analogue scale can also be used to assess the same question, with the ability to have scores which
are not whole numbers as the above mentioned scale is (211). The 1-5 scale is the most
commonly utilized assessment for determining TKR patient satisfaction and subsequent division
into different groupings (i.e. very satisfied and satisfied vs dissatisfied and very dissatisfied).
Some measurement tools have chosen to elaborate on this construct with questions
relating to satisfaction with particular activities as a means of gauging a possible change in
satisfaction with tasks of varying difficulty. One researcher asks an additional question to the
original satisfaction question of whether the patient is satisfied with his/her level of activity after
the TKR procedure (212). Another researcher furthered the investigation of the dissatisfied
patients with follow up questions of the symptoms or functional disabilities which may have led
to the dissatisfaction as a means of trying to get to the root cause of dissatisfaction (9). As
mentioned earlier, the FJS creators believe the ultimate level of satisfaction is the ability to forget
the replaced joint exists in everyday life (158). This is a construct of the authors though and was
never actually used to subdivide a population into satisfied and dissatisfied groups. A threshold
score for the tool would need to be set to deem a patient satisfied/dissatisfied. It could then be
statistically analyzed against an additional determinant of satisfaction to see its effect. In
general, every measure of satisfaction reverts back to the simple question of asking patients to
rate their overall satisfaction with the TKR. While this is a subjective assessment, it is an
essential query as patient satisfaction is the ultimate goal of TKR. A more sophisticated
analysis, such as a logistic regression, may provide additional insight into this complex construct.
The terms “satisfaction” and “success” are often used interchangeable, but the two
constructs are substantially different. In reference to a total knee replacement (TKR) population,
94
“satisfaction” is a patient-based term while “success” is most appropriately determined by
medical personnel (surgeon, doctor, physical therapist, etc.). “Satisfaction” has been defined by
a subjective question posed to patients, “Are you satisfied with your knee replacement?”
Patients are given a four or five point Likert scale (different researchers have used different
scales as there is no standardized scale) with responses gauged between very dissatisfied and
very satisfied (17, 202). Conversely, “success” for an operation has been defined by the
reduction of pain, restoration of joint flexion range of motion, improved joint alignment, and a
restored ability to perform activities of daily living, as is evident in the Knee Society System
scoring tool (5). There have been several instances where the patients and the medical personnel
do not agree on the success of the TKR procedure, with the medical personnel often having a
more positive outlook on the outcome of the procedure than the patient population (13, 195,
210). The surgeons will typically base their evaluation of the success of a TKR procedure on a
multitude of physical aspects of the procedure, such as range of motion (ROM) of the knee post-
surgery, ability to perform physical activities, and the surgical alignment of the knee (as in a
correction of an alignment deformation from prior to the surgery). Patients will evaluate their
satisfaction based on the surgery meeting their pre-surgery expectations, pain relief, and the
ability to return to activity post-surgery (13). Patients hope for an ability to return to the
functioning they had prior to the onset of their knee issues. When making this assessment, some
expectations for the operation become unrealistic as a passage of time and neglect of fitness
levels may impair a patient’s abilities. Surgeons take this into account when evaluating success
of the surgery (13). A 70 year old patient who has been neglecting their fitness since the onset of
knee pain 15 years earlier may not be able to return to the same level of activity as prior to the
onset of knee pain.
95
Like the surgeons, patients also assess the physical aspects of the procedure, but mostly
based on their expectations of how the procedure will affect their physical function abilities and
pain levels. The total level of pain and function does not seem to be as important to the patient
as the relative reduction in overall pain post-surgery (18). There can be additional factors
contributing to satisfaction levels for the patients that are not directly related to the surgery site
(i.e. pain relief in the knee, ability of the knee to function normally), such as the need for outside
assistance (help completing certain tasks), lack of social interaction or a social support network,
or the presence of comorbidities, all of which have been shown to reduce satisfaction levels (17).
These factors do not necessarily affect the determination of good clinical outcomes (success) for
the medical team but may factor into the patient’s level of satisfaction. Given that the patients
and medical teams have different criteria for judging the procedure, there is the possibility of
disagreements between the two. There have been reports of good clinical outcomes with
dissatisfied patients as well as bad clinical outcomes with satisfied patients (17), illustrating the
discontinuity between the patients and the medical personnel with respect to the evaluation.
An interesting, albeit rarely used, construct for assessing patient satisfaction is the ability
to forget about the replaced joint during everyday life (158). It could be argued that the ability to
forget about the presence of an artificial joint may mean that pain is minimal to non-existent and
that functional limitations are the same, thus adding to satisfaction. At the same time, it could be
argued that the patient is merely coping with the pain or functional limitations and found a way
to deal with a continuing presence of them through compensatory strategies or mindsets which
reduce the experience of discomfort or dissatisfaction. This examination of proprioception has
been suggested as the “ultimate goal in joint arthroplasty” (158), but it should be combined with
assessing the above mentioned factors.
96
Psychological Factors of Satisfaction
While the question for determining satisfaction is relatively simple in that it is a single
question assessing the global outcome, in reality many factors influence satisfaction beyond the
physical components traditionally investigated in the TKR population. In a comparison of
psychological tools to the WOMAC and FJS, 54.3% of the variance in the WOMAC scoring and
30% of the FJS scoring were explained in a multivariate regression model using two
psychological measurement tools: the Catastrophizing scale and the Brief Symptom Inventory.
This led to the conclusion that there is a strong relationship between psychological status and
orthopedic outcomes (213). The WOMAC total score had a high correlation with the
Catastrophizing scale (up to a correlation value of 0.79). The WOMAC subscales are not
designed to represent any psychological component, but based on the correlations, there are
clearly psychological factors which may be related to the physical components (213). Based on
content alone, the WOMAC is a physical tool, but there seems to be underlying psychological
factors which contribute to the scoring within those physical dimensions, based on correlations
between the two tools. The FJS, on the other hand, while being largely deemed a physical tool,
the “joint awareness” it references could easily be argued as possessing a psychological construct
in that it deals with awareness of a feeling of the knee replacement. The tool measures the
awareness of a patient’s knee replacement, thereby asking them to assess their proprioception of
the area of the knee, where the prosthetic exists.
In further enhancing the psychological construct, types of joints replaced were compared
to the psychological components (Catastrophizing Scale and Brief Symptom Inventory). The
psychological components had stronger associations than the location of the joint replacement,
suggesting that the idea behind surgery is more important than the surgery itself (213). Having a
97
joint replaced is a major operation which requires some mental skills to deal with. Injury and the
subsequent recovery is certainly a physical event; however it is also a psychological and
emotional event (214). The recovery can be long and very defeating if progress does not occur
as quickly as anticipated, leading to a decreased level of satisfaction as the patient may not be
able to return to the physical abilities as previously anticipated (111). Successful coping
behaviors (such as positive attitude, positive outlook on treatment, rehabilitation adherence) have
been instrumental in aiding in the recovery process while unsuccessful coping behaviors hinder
recovery (poor adherence, poor attitude) (215). It should be noted that the relationship between
orthopedic measures and psychological measures may suggest poor divergent validity, which can
impair the accurate assessment of orthopedic outcomes (213).
Patient Satisfaction for TKR Procedures
Satisfaction is a complex construct which could be argued as a psychological construct
rooted in physical properties, especially in pertaining to the TKR population. The vast majority
of the studies examining satisfaction in the TKR population have done so with a simple question
used to divide the population: “How satisfied are you with your knee replacement?”. The
patients are then provided with a 5 point scale which corresponds to options of very satisfied (5),
somewhat satisfied (4), neutral (3), somewhat dissatisfied (2), and very dissatisfied (1), possibly
with some slight wording deviations but still arriving at the ability to be able to distinguish
between satisfied and dissatisfied patients (9, 11, 17, 18, 111, 192, 194, 203, 210). A visual
analogue scale can also be used to assess the same question, with the ability to have scores which
are not whole numbers as the above mentioned scale is (211). The 1-5 scale is the most
commonly utilized assessment for determining TKR patient satisfaction and subsequent division
into different groupings (i.e. very satisfied and satisfied vs dissatisfied and very dissatisfied).
98
Some measurement tools have chosen to elaborate on this construct with questions
relating to satisfaction with particular activities as a means of gauging a possible change in
satisfaction with tasks of varying difficulty. One researcher asks an additional question to the
original satisfaction question of whether the patient is satisfied with his/her level of activity after
the TKR procedure (212). Another researcher furthered the investigation of the dissatisfied
patients with follow up questions of the symptoms or functional disabilities which may have led
to the dissatisfaction as a means of trying to get to the root cause of dissatisfaction (9). The
Forgotten Joint Score creators believe the ultimate level of satisfaction is the ability to forget the
replaced joint exists in everyday life (158). This is a construct of the authors’ though and was
never actually used to subdivide a population into satisfied and dissatisfied groups. A threshold
score for the tool would need to be set to deem a patient satisfied/dissatisfied. It could then be
statistically analyzed against an additional determinant of satisfaction to see its effect. In
general, every measure of satisfaction reverts back to the simple question of asking patients to
rate their overall satisfaction with the TKR. While this is a subjective assessment, it is an
essential query as patient satisfaction is the ultimate goal of TKR.
An examination of the satisfaction results for TKA patients has yielded some wide ranges
of responses. The rate of dissatisfaction ranges between 6 and 19%, neutral from 2 to 12%, and
satisfied from 70-89% (9, 11, 17, 18, 111, 192, 194, 203, 210, 216). Best case scenario is an
89% satisfaction rate which still leaves 11% of the population in a neutral or dissatisfied state,
which is a significant portion of the population. A visual analogue scale measurement of
satisfaction revealed 68% of the people sampled had an 80/100 mm score or better (211), which
still leaves a lot to be desired for increasing the satisfaction level. These outcomes will,
however, differ with respect to time as short term and long term time periods have been shown to
99
be different with respect to improvement relating to satisfaction (217). Additionally, the type of
procedure done with respect to knee replacements plays a role as TKR patients have been shown
to have higher dissatisfaction rates compared to bicompartmental knee replacement patients
when the ACL is retained (218).
Patient satisfaction with TKA procedures have been shown to have significant
correlations with improved SF-36 scores, WOMAC pain scores, improved knee function
(assessed through self-evaluation using a functional outcome assessment questionnaire), and
overall physical function (assessed with same functional outcome questionnaire) (17). The
dissatisfied patients have also been shown to have decreased WOMAC pain, stiffness, and
function scores (meaning the patients experienced an increase in pain and stiffness and a
decrease in function), and decreased SF-36 physical functioning, role physical, pain, vitality,
social function, and role emotional scores (9). The OKS scores also improved with patient
satisfaction, showing a worse pain score for dissatisfied patients and a strong correlation between
pain and function (18, 192). The visual analogue scale had reported that 84% of patients
surveyed had pain scores less than 20/100 mm (211). In reporting correlations of satisfaction
with specific measurement tools, the OKS reported the highest correlation value followed by the
WOMAC and then the SF-36 and SF-12 physical component summary and then mental
component summary (194). In a study examining both pre-operative and post-operative
measurements, the only pre-operative measurement which was shown to be related to
dissatisfaction was self-reported mental functioning (219), illustrating that state of mind prior to
the procedure is crucial. A good state of mind should be obtained prior to committing to the
procedure given the dedication needed to successfully rehabilitate. This same group showed
100
increased 6 minute walk test times, increased pain, increased physical limitations, and reduced
WOMAC and SF-36 scores (219).
In an examination of demographic components, men were more satisfied with the
procedure than women and older patients were more satisfied than younger patients (18). It has
been reported that 7% of those older than 70 were dissatisfied while 10% of those under 70 were
dissatisfied (212). An additional study reported the highest dissatisfaction (19%) was for
patients in the age group of 60-75 (111). Yet another study rated the highest dissatisfaction
(19%) with those patients under the age of 55 (220). This is to be expected as the younger the
individual, the likely the higher desire to return to a more active lifestyle than the older adults.
Consequently, the assessment of highest dissatisfaction being within the 60-75 age group is
likely because this population is still young enough to desire to return to a more active lifestyle,
but not as capable of returning to an active lifestyle as the population under 60, based on physical
components of healing and strength. Age can play a large role in recovery as well as perceptions
of what their physical abilities should be. It has been suggested though that the age of a patient
is not nearly as important as their physiological age, as a 70 year old may be younger
physiologically than a 50 year old based on how well their body was taken care of (221).
The dissatisfied population has a decreased desire to have the procedure done again
(meaning if presented with the option of a TKR procedure, they would not have the surgery
done) compared to the satisfied (192). There is no single underlying construct to this perception,
and may actually be a combination of both physical and mental components of the procedure.
This may suggest that the dissatisfied patients’ lives were more satisfying prior to the surgery,
although this has yet to be confirmed. Different replacement types also showed different levels
of satisfaction, with cemented TKAs being more satisfied than unicompartmental procedures
101
(18). However, no satisfaction differences were present when examining the different types of
implants (such as posterior stabilizing or posterior cruciate ligament retaining) or whether there
was patellar resurfacing (203, 210).
Aside from overall satisfaction with the TKR, a few studies have assessed patient’s
satisfaction with specific situations related to the TKR. A few of the assessed components did
have differences with satisfaction. Satisfaction for pain relief varied between 72-86%. There
was 70% satisfaction with pain relief during stair negotiation, 85% during walking, and 84%
during sitting (210). Higher pain has been associated with lower satisfaction levels (222). Knee
stiffness or the use of analgesics at least once a week have been associated with dissatisfaction
(111). Satisfaction also largely varied between functional activities, ranging from 70-84%.
There was 86% satisfaction with walking on a flat surface, 70% satisfaction with getting out of a
car, 73% with ascending stairs, 82% with getting out of bed, 84% with lying in bed, and 83%
with household chores (210). Over half of the dissatisfied patients are not as active as they
thought they would be, 32% are less active after surgery than they were before, 53% could not
perform the activities they desired to, and 71% reported some difficulty with ADLs (111). The
dissatisfaction with given tasks reflects the dissatisfied patient groups’ difficulty with certain
tasks. Dissatisfied groups have been shown to have increased difficulty with kneeling,
recreational activities, siting/rising from a chair, stairs, working, and walking, often requiring
them to employ a gait pattern with a limp (223). Activities with high flexion have been rated as
more important to dissatisfied groups, likely due to their inability to perform them, leading to a
desire for the ability to be able to successfully perform high flexion activities without limitation
or pain (223, 224). The more capable a person is of participating in ADLs and athletic activities,
the higher the satisfaction rates. In order to achieve this, a behavioral contract has been
102
suggested as means of implementing weight loss and regular exercise as they have been shown
to improve a person’s ability to perform ADLs (225). The contract serves as a more formal
commitment to the activity program, thereby increasing participation.
Even with patients with an unlimited ability to walk and climb stairs, 17% were still
unsatisfied, suggesting that additional underlying factors outside of physical limitations may be
contributing to dissatisfaction (11). From that same data set, 72% of sedentary patients who
could only walk five minutes or less were still satisfied (11). This strongly suggests factors
outside of physical limitations play significant roles in determining patient satisfaction,
warranting examinations outside of just the physical components when investigating patient
satisfaction with TKA procedures. This statement is not made to suggest that physical
limitations do not play a role in satisfaction, as they clearly do. Dissatisfied patients have been
shown to have decreased knee flexion and ROM at the knee (9), which can obviously hinder
physical performance and possibly contribute to pain. In a study which split TKR patients based
on flexion ROM levels, the high flexion group showed no dissatisfaction while the midflexion
group showed a 17% dissatisfaction rate and the low flexion group showed 16% dissatisfaction.
However, the magnitude of the maximum degree of knee flexion did not affect satisfaction, only
the total ROM (226). Moderate correlations have been shown between functional ROM and
patient satisfaction, leading some to conclude that functional outcome is not necessarily an
indicator of patient satisfaction (227) and that single factors are likely not going to explain
satisfaction (226). One study has shown a decrease in ROM by 11° which was coupled with
increased anxiety, depression, and pain, but interestingly, there were no differences in the 6
minute walk test, chair stand test, knee laxity, patellar tenderness, or mechanical axis angles
when compared to the satisfied TKR patients (16). However, 48% of the variance in satisfaction
103
can be explained by the mental health, emotional role function, and social function subscales in
the SF-36 (17). In additional research, 97% of the variance on patient satisfaction is related to
the meeting of the pre-op expectations, achievement of satisfactory pain relief, the patient’s
hospital experience (subjective), their pre-op OKS score, and their 12-month physical status
post-op (192). Other reports for the greatest predictors of dissatisfaction were the presence of
other joint problems, increased pain at 2 years post-operatively, and greater functional limitations
(228). It is evident that examination of the physical components is not enough to truly derive an
accurate assessment of patient satisfaction and the underlying causes of dissatisfaction.
Post-operative walking limitations for TKR patients were predicted by pre-operative
walking limitations, high BMI, lower gait speed at 1 month post-operatively, the presence of
contralateral knee pain, and the use of a quadstick (a four-pronged walking cane which is more
stable than a regular, single prong cane) pre-operatively (229). This helps to illustrate the
complexity of satisfaction as a multitude of factors play into the outcomes for the patients. An
additional consideration is management of expectations for the patients. Unrealistic expectations
of recovery or outcome of the procedure have been shown to decrease satisfaction levels (221,
230). At the same time, while managing expectations, dissatisfaction can increase despite no
significantly noticeable functional limitations (231). Patients who reported higher expectations
for the outcome of the operation have shown 4.3% dissatisfaction rates, but more importantly,
37% neutral rates, meaning neither satisfied nor dissatisfied (202). While a neutral outlook can
be argued as better than dissatisfied, it still does not meet expectations for the procedure. The
measure of satisfaction is subjective for the patient and largely based on their own perception of
their situation. How a patient rates their satisfaction may partially depend on their overall mood
as well. In a comparison of two different icing treatment groups, satisfaction was seen to be
104
higher for a group using a cryopneumatic cooling device compared to a group using traditional
ice treatments, but both groups reported being satisfied (231). This illustrates that is may not
make sense to disseminate between “satisfied” and “very satisfied” patients as there are no
qualifying criteria to separate the two responses. The majority of studies place these two
responses in a group together (192, 203, 226), which seems to be the more appropriate
classification than splitting them up.
The research on the dissatisfaction has provided a substantial amount of information,
however there are some areas which need further examination. First, there is minimal
biomechanical data on the dissatisfied population. There is a lack of research on the
biomechanics of gait and stair climbing for the TKR population which have reported on patient
satisfaction. Kinematics and kinetics play a large part on the evaluation of TKR patients, yet
there has been relatively little information provided on the kinematics and kinetics of the
dissatisfied population and how they compare to satisfied TKR patients and healthy controls.
Additionally, it may be of interest to note how biomechanical variables relate to strength,
balance, and functional tasks for these same populations.
Closing Statement
The current literature review illustrates some of the current research examining the TKR
population with respect to biomechanics, strength, balance, and survey data. The research in the
TKR field has been continuously expanding, however there is currently a lack of research on the
dissatisfied patient population. Research on biomechanical gait profiles, strength, and balance
needs to be conducted on the dissatisfied population for the purpose of identifying areas of
deficit which could be addressed to potentially improve patient outcomes and therefore patient
satisfaction with the TKR procedure. Additionally, information obtained through this research
105
may provide information to help improve future TKR implant designs, which could potentially
improve patient satisfaction. A more complete biomechanical profile assessment of the
dissatisfied population may provide opportunities to reduce the percentage of dissatisfied
patients by providing information which is addressable in regards to how a patient performs in
various testing scenarios.
106
CHAPTER III
OVERGROUND WALKING BIOMECHANICS OF SATISFIED AND DISSATISFIED
TOTAL KNEE REPLACEMENT PATIENTS
107
Abstract
Patient dissatisfaction after total knee replacement (TKR) procedures is likely influenced
by both subjective and objective aspects. Increased pain and reduced performance on clinical
tests have been shown in the dissatisfied patients, however, it is unknown how overground
walking kinematics and kinetics are different between dissatisfied and satisfied TKR patients.
Therefore, the purpose of this study was to compare overground walking lower extremity
kinematics and kinetics of dissatisfied TKR patients to satisfied patients and healthy controls.
Nine dissatisfied TKR patients, fifteen satisfied TKR patients, and fifteen healthy controls
performed overground walking trials. A 2 x 3 repeated measures ANOVA was used to assess
differences between groups and limbs (p<0.05). The dissatisfied patients showed reduced 1st and
2nd peak VGRFs, flexion ROM, loading-response extension moments, and loading-response
abduction moments compared to healthy controls. First and 2nd peak VGRFs and flexion ROM
were reduced in the replaced limb of the dissatisfied patients compared to their non-replaced
limb. Push-off plantarflexion moments were reduced in the dissatisfied patients compared to
healthy controls and satisfied patients. Dissatisfied patients also reported increased knee joint
pain and reduced preferred gait speed. Limb asymmetry was evident in the dissatisfied patients
for VGRF and knee extension moments. Future research should examine methods to address the
asymmetrical loading as a potential means to improve patient satisfaction rates.
Keywords: total knee replacement, arthroplasty, satisfaction, overground walking
108
Introduction
Overall, most total knee replacement (TKR) operations are considered successful as there
are often reductions in pain levels and improvements in range of motion (ROM) in the replaced
knee (6-9). Patient satisfaction rates for the procedure have been reported between 81-97% (10,
11). This leaves a significant portion of the TKR population as dissatisfied with the outcomes of
the replaced knee. Post-operative pain (12) and functional limitations (13) are commonly
reported by dissatisfied patients. This often results in decreased performance in common clinical
tests (such as timed-up–and-go or sit to stand test) when compared to healthy controls (14, 15).
These tests are often seen as a defining point for “success” of the operation. However, these test
results do not sufficiently explain why the TKR patients are dissatisfied with the TKR outcomes.
Examination of walking biomechanics may help to provide quantifiable and detail insights into
their dissatisfaction.
There have been several studies investigating the TKR population as a whole compared
to healthy controls. These studies have examined overground walking, with special focus on gait
velocity (20, 21), sagittal plane knee ROM (20, 22), and frontal plane knee moments (20, 21).
Some research has shown a return to vertical ground reaction force (VGRF) symmetry between
the replaced and non-replaced limb in TKR patients (46), a knee flexion ROM comparable to
healthy controls (25, 26), and similar peak knee abduction moments in replaced limbs compared
to healthy controls (26). However, not all studies are in agreement with these results as
differences have been shown with respect to VGRF, ROM, and joint moments (24, 26, 46).
Reduced peak knee extension moments have been shown in TKR patients (46), likely due to a
reduced quadriceps strength, resulting in a quadriceps avoidance gait pattern. This reduction in
knee extension moments sometimes results in an increase in knee abduction moments (38, 114,
109
138). These changes in joint moments may indicate a reduced level of recovery given the
differences compared to healthy controls, which may be further amplified in the dissatisfied
patient population, although this has yet to be examined.
Patient dissatisfaction is likely influenced by multiple factors, including both subjective
and objective aspects. Issues with pain, perceived function, and physical abilities may all
contribute to dissatisfaction. Currently, patient dissatisfaction has been largely quantified based
on survey and functional test data with very little information with respect to the three-
dimensional kinematics and kinetics of TKR patients, including overground walking, a simple
but necessary daily task. How dissatisfied patients differ from satisfied patients and healthy
controls in overground walking mechanics is largely unknown. Therefore, the purpose of this
study was to compare the lower extremity kinematics and kinetics of dissatisfied TKR patients to
satisfied TKR patients and healthy controls during overground walking. It was hypothesized that
dissatisfied TKR patients would exhibit reduced knee extension moments and increased knee
abduction moments in their replaced limb compared to their non-replaced limb in level walking.
Additionally, it was hypothesized that dissatisfied TKR patients would exhibit increased reduced
knee extension moments and knee abduction moments compared to satisfied TKR patients and
healthy controls in level walking.
Materials and Methods
Participants
Twenty four TKR participants were recruited from a local orthopaedic clinic. There were
nine dissatisfied TKR patients (68.0±4.2 years, 1.69±0.07m, 81.0±18.6 kg, 34.6±14.3 months
post-surgery) and 15 satisfied TKR patients (66.6±6.3 years, 1.76±0.10m, 90.2±17.0 kg,
29.3±12.8 months post-surgery). The inclusion criteria for TKR patients were the presence of a
110
unilateral total knee replacement performed by a single surgeon, at least 12 months but less than
60 months post-operative, and between the ages of 50 and 75. Exclusion criteria were any
additional lower extremity joint replacements, any additional diagnosed osteoarthritis of the hip,
knee, or ankle, BMI greater than 38 or neurological diseases. TKR patients were then asked,
“How satisfied are you with your total knee replacement?” The available responses were “very
dissatisfied, dissatisfied, neutral, satisfied, or very satisfied.” “Neutral” responses were
excluded. “Very dissatisfied” or “dissatisfied” was placed into the Dissatisfied group and
“satisfied” or “very satisfied” were placed into the Satisfied group. Fifteen healthy controls
(60.7±9.2 years, 1.75±0.09m, 77.7±11.8 kg) were recruited with the same exclusion criteria as
the TKR groups.
Instrumentation
A twelve-camera motion analysis system (240 Hz, Vicon Motion Analysis Inc., Oxford,
UK) was used to obtain three-dimensional (3D) kinematics during testing. All participants wore
standardized running shoes (Noveto, Adidas, Herzogenaurach, Germany). Anatomical
retroreflective markers were placed bilaterally on the acromion processes, iliac crests, greater
trochanters, medial and lateral femoral epicondyles, medial and lateral malleoli, 1st and 5th
metatarsal heads, and 2nd toe. A cluster of four retroreflective markers on a semi-rigid
thermoplastic shell was placed on the lateral aspects of both shanks and thighs, and on the
posterior aspect of the pelvis and the thoracic cage, attached via an elastic neoprene wrap with
hooks and loops of Velcro. Four individual tracking markers were placed on the heel counter of
the shoe. Anatomical and tracking markers were kept on for the static trial and anatomical
markers were removed prior to testing trials. One force platform (1200 Hz, BP600600,
American Mechanical Technology Inc., Watertown, MA, USA) was used to measure the ground
111
reaction force (GRF) and moments of forces. Walking speed during all trials was monitored by
two sets of photo cells (63501 IR, Lafayette Instrument Inc., IN, USA) and electronic timers
(54035A, Lafayette Instrument Inc., IN, USA).
Experimental Procedures
All participants signed an informed consent form, and completed a physical activity
readiness survey (PAR-Q) to assess cardiovascular risks to exercise and a Western Ontario and
McMaster Universities Osteoarthritis Index (WOMAC) for both knees (178). TKR participants
also completed a Forgotten Joint Score (FJS, (158)). Following completion of the surveys, all
participants completed a five-minute warm up on a treadmill, walking at a self-selected speed.
Participants were then fitted with the retroreflective markers. Participants performed three
practice walking trials over a 10 meter runway at a self-selected speed. The practice trials were
to familiarize themselves with the lab setup, find a starting position to ensure proper foot contact
with the force platform without stutter stepping or targeting, and determine average walking
speed. A speed range (mean speed ± 5%) was used to control participant speed during
experimental trials. All participants performed five trials in two randomized conditions, one
with the left foot in contact with the force plate and one with the right foot. A trial was discarded
and repeated if the speed range was not met, the foot was not in full contact within the bounds of
the force platform, or targeting occurred. A numerical visual analogue pain (VAS) scale was
used to assess pain level in both knees for TKR and healthy participants at the end of the test
conditions.
Data Analyses
Visual3D biomechanical analysis software suite (version 5.0, C-Motion, Inc.,
Germantown, MD, USA) was used for 3D kinematic and kinetic variable computations for
112
overground walking data. A Cardan rotational sequence (X-y-z) was used for 3D angular
kinematics computations and the conventions of the angular kinematic and kinetic variables were
defined using the right-hand rule. Positive values indicate ankle dorsiflexion, inversion, and
internal rotation, knee extension, adduction, and internal rotation, and hip flexion, adduction, and
internal rotation angles and joint moments (computed as internal moments). Kinematic and GRF
data were filtered using a fourth-order low-pass Butterworth filter with a cut-off frequency of
8Hz before kinematics and joint moment calculations. Raw GRF were filtered separately using a
fourth-order low-pass Butterworth filter with a cut-off frequency of 50 Hz to calculate GRF
variables. Critical events and values, including peak VGRF, knee extension, abduction, and
internal rotation moments, loading-response knee flexion, adduction, and external rotation ROM,
and sagittal plane hip and ankle ROM and moments, were chosen using customized computer
programs (VB_V3D and VB_Table, MS Visual Basic 6.0, USA). Joint moments represent
internally applied moments, were reported in the proximal reference system and were normalized
to body mass (Nm/kg). GRF variables were normalized to body weight (BW). Averages across
the five trials of selected variables for every condition for each participant were used in statistical
analyses.
Statistical Analyses
A 2 x 3 (limb x group) mixed model analysis of variance (ANOVA, p<0.05) was
performed to detect differences between limbs and groups for kinematic and kinetic variables
(Version 9.4, SAS, Cary, NC, USA). A one-way ANOVA (p<0.05) was performed on
demographic and survey data to test for differences between the three groups. When the
ANOVA revealed a significant interaction or main effect, post-hoc comparisons for multiple
pairwise comparisons were used to compare means between limbs and groups. Post-hoc
113
comparisons were made using adjusted p values of 0.00625 for 2x3 ANOVA significant
interactions, 0.0167 for 1x3 ANOVA comparisons and group main effects, and 0.05 for limb
main effects. Post hoc comparisons were only made between the TKR replaced limbs against
healthy dominant limbs and TKR non-replaced limbs against the healthy non-dominant limbs,
under the assumption that any differences between the dominant and non-dominant limbs are
random and small when present in the healthy participants.
Results
There were significant differences (p=0.0034) in age for the dissatisfied and satisfied
group groups compared to the healthy group. Walking speed for the walking trials revealed a
significant difference between the groups, with the dissatisfied group walking slower than both
satisfied and healthy groups.
Significant interactions were present for 1st and 2nd peak VGRF (Table 1). Both the
dissatisfied and satisfied groups had reduced 1st peak VGRF in their replaced limb compared to
their non-replaced limb and the healthy group (p<0.0060 for both comparisons). The 2nd peak
VGRF was reduced in the replaced limb of the dissatisfied group compared to their non-replaced
limb and both satisfied and healthy groups (p<0.0021 for all comparisons). A significant
interaction was present for knee flexion loading ROM, which was significantly reduced for both
dissatisfied and satisfied groups in their replaced limbs compared to their non-replaced limbs and
healthy participants (p<0.0061 for all comparisons). A significant interaction was present for
peak knee loading-response abduction moments, with reduced moments for the dissatisfied
patients in their replaced limb compared to the healthy group (p=0.0061). Transverse plane knee
ROM showed a significant interaction, exhibiting a reduction in the replaced limb of the
dissatisfied group compared to their non-replaced limb (p=0.0040). Both peak knee loading-
114
response internal rotation and push-off external rotation moments showed significant
interactions. The dissatisfied patients showed reduced peak internal rotation moments in their
replaced limb compared to their non-replaced limb and both satisfied and healthy groups
(p<0.0058 for all comparisons). Push-off external rotation moments were reduced for the
replaced limbs of both TKR groups compared to the healthy group (p<0.0023 for all
comparisons). Significant limb main effects were present for knee extension loading moments
and push-off moments, with reductions in the replaced limbs compared to the non-replaced limbs
(p < 0.0317for all comparisons, Table 1). Significant group main effects were also present for
peak knee extension loading moment and loading-response knee adduction ROM. The
dissatisfied group showed reductions in peak loading-response extension moments compared to
the healthy group (p=0.0052). The satisfied group showed reduced loading-response knee
adduction ROM compared to the healthy group (p=0.0139).
A significant interaction was present for ankle dorsiflexion ROM (Table 2). Both
dissatisfied and satisfied groups had increased ankle dorsiflexion ROM in their replaced limbs
compared to their non-replaced limbs as well as reductions compared to the healthy group
(p<0.0032 for all comparisons). Significant limb main effects were present for both peak
loading-response dorsiflexion and push-off plantarflexion moments. Peak loading-response
dorsiflexion moments were higher in the replaced limb compared to the non-replaced limb
(p=0.0287) while push-off plantarflexion moments were lower in the replaced limb compared to
the non-replaced limb (p=0.0110). A significant group main effect was present for the ankle
push-off plantarflexion moment, with reduced moments in the dissatisfied group compared to the
satisfied (p=0.0005) and healthy (p=0.0006) groups.
115
Significant interactions were present for the WOMAC total and all subscale scores
(p<0.0007 for all tests, Table 3). The dissatisfied group showed increased scores for the total
and all subscale scores compared to their non-replaced limb and the satisfied and healthy groups.
Significant interactions for passive knee flexion ROM revealed reduced ROM for the replaced
limb compared to the non-replaced limb for both the dissatisfied and satisfied groups, with
additional reductions for both groups compared to the healthy group (p<0.0041 for all
comparisons). Additional significant interactions were present for all pain measurements
(p<0.0013 for all tests), with increased pain for the dissatisfied group in their replaced limb
compared to their non-replaced limb and both satisfied and healthy groups.
Discussion
The purpose of this study was to examine lower extremity kinematics and kinetics of
dissatisfied TKR patients during overground walking in comparison to satisfied TKR patients
and healthy participants. The first hypothesis was that the dissatisfied patients would exhibit
decreased knee extension moments and increased knee abduction moments in their replaced limb
compared to their non-replaced limb. The results of this study were in partial agreement with
our first hypothesis. Interactions and post-hoc comparisons did not show any differences
between the replaced and non-replaced limbs for knee extension and abduction moments.
However, there was a significant limb main effect for the loading-response and push-off knee
extension moments, with replaced limb’s being lower than the non-replaced limbs. Some studies
have reported increased peak knee extension moments (external flexion moments) in healthy
controls compared to TKR patients during the loading-response phase (24, 26) while others have
reported no difference between patients and healthy controls (20, 25). In this study, there was a
significant main group and limb effect for the loading response extension moment, whereby the
116
dissatisfied group had lower extension moments than the satisfied and healthy group and the
replaced limb had lower moments than the non-replaced limb. It has been suggested that these
reductions are due to a quadriceps avoidance gait pattern and that targeted therapy is necessary to
remedy this (25). It should be of note though, that the bi-modal knee extension pattern are more
common post-operatively compared to pre-operatively (53), indicating a return to normal
movement patterns. In this study, all patients except for one per group displayed a bi-modal
knee extension pattern. Despite reduced values, this could be considered a step in the right
direction for rehabilitation. Increased quadriceps strength may help with improving ability to
handle increased joint loads, although it is unlikely the replaced knee will ever function as well
as a non-replaced knee. The dissatisfied group showed decreases in quadriceps strength in their
replaced limb compared to their non-replaced limb (Chapter V), suggesting their reduced ability
to handle the knee joint loading during gait.
The peak knee extension moment during stance phase has been shown to be reduced in
TKR patients compared to healthy controls, likely due to reductions in quadriceps strength and
therefore an avoidance pattern, resulting in the hip compensating for the knee (46). This study
did not support this hip compensation strategy as no hip extension moments or extension ROM
differences were present in either patient group. However, there were differences with respect to
sagittal ankle patterns. The dissatisfied and satisfied groups both showed increased ankle
dorsiflexion ROM during loading response in their replaced limbs compared to their non-
replaced limbs and the healthy group. This may be linked to the reduced knee flexion ROM in
both groups. Yet, the satisfied group was able to achieve gait velocity similar to the healthy
controls while the dissatisfied group was not.
117
The flexion ROM during loading was significantly reduced in both dissatisfied and
satisfied replaced limbs compared to non-replaced limbs and healthy limbs. This may be in part
due to reduced gait velocity and therefore reduced need for knee flexion for dissatisfied patients,
but does not hold true for the satisfied patients. Similar amounts of knee flexion ROM have been
reported between TKR patients and healthy controls (25, 26), which is not in agreement with this
study. This reduction has been seen to carry on for the first year after surgery, with a reduced
dependence on the operated knee having been shown at 12 months post-operatively compared to
the non-operated limb and healthy controls (46), which still exists for both patient groups in this
study at an average of over two years post-operation.
Initially, it should be noted that the dissatisfied group walked with a reduced gait speed
compared to the other two groups, averaging 1.16 m/s. This is close to previously reported pre-
operative values of 1.13 m/s for TKR patients (25). The reduced speed could be due to the
increased pain levels experienced by the dissatisfied group in their replaced limb compared to all
other groups, which subsequently has an effect on their level walking profiles. Additional
studies have reported no differences in gait velocity at 12-18 months post-operation for TKR
patients compared to healthy participants, with velocity being retained at an average of 46
months after surgery (22, 26). This held true for our satisfied TKR participants at an average of
29.3 months after surgery (with no gait velocity differences compared to healthy participants) but
not for the dissatisfied patients at 34.6 months after surgery. With this reduction in speed and
increase in pain came several other mechanical changes for the dissatisfied patients. An
additional decrease in peak push-off plantarflexion moments was present in the dissatisfied
group, which may explain the reduced gait velocity. The impairment of the knee extensors also
118
appeared to negatively affect the plantarflexors, which should be further examined in future
research.
The second hypothesis, which hypothesized reduced knee extension moments and
increased knee abduction moments for the dissatisfied patients compared to the satisfied and
healthy participants, was partially confirmed. The loading response knee extension moment
showed a group main effect, whereby the dissatisfied group showed reduced moments compared
to the satisfied and healthy groups. Additionally the dissatisfied group showed a reduced
loading-response knee abduction moment in their replaced limb compared to the dominant limb
of the healthy group.
There are mixed results with respect to peak knee abduction moments during gait in the
literature. Some research has shown no difference between TKR patients and healthy controls
(26), reduced loading-response and push-off peaks (20), and increased peaks in non-replaced
limbs compared to replaced limbs but no difference between the replaced limb and healthy
controls (38, 114, 138). In this study, only the dissatisfied replaced-limb showed lower peak
loading-response abduction moments than the healthy participants, with no other differences
evident. This is partially in agreement with previously mentioned research with reduced loading-
response peaks (20). Other research has shown at an average of 2 years post-operation, TKR
patients showed increased peak abduction moments in their replaced limb compared to healthy
controls, but no difference between the replaced and non-replaced limbs or the non-replaced and
controls (19). No gait velocity differences were evident, however the knee extension moment
was lower in the replaced limb compared to the healthy controls (19). It does illustrate that the
TKR patients may undergo a compensatory transformation whereby the load is transferred to
another plane. The reduced knee extension moment may have been transferred to the frontal
119
plane. This was not the case for the dissatisfied patients as the reduced gait velocity was
experienced along with reduced knee extension moments. This may be indicative of avoiding
increased pain associated with increases in frontal-plane medial knee loading.
Asymmetry was a problem for the dissatisfied participants, indicating a failure to perform
equally on both limbs. For both 1st and 2nd peak VGRF, the dissatisfied group showed reduced
forces on their replaced limb compared to their non-replaced as well as compared to the healthy
group. The satisfied group also showed the similar imbalance in their 1st peak VGRF. A return
to VGRF symmetry between the two limbs has been previously reported in TKR patients by 12
months after surgery (46), however this was not the case for either patient group in this study. .
Asymmetry in the movement pattern often indicates increased dependence on one limb, which
can prove problematic as this could lead to the contralateral knee needing replacement, as
happens in approximately 40% of TKR patients within ten years (118). A reduction in VGRF in
the replaced limb may lead to the non-replaced limb bearing increased overall loading to the
body, which can lead to long-term joint health issues. The changes in VGRF can also impact the
joint moments, as is evidenced in this study where simultaneous VGRF and knee extension
moments were seen for the dissatisfied patients in their replaced limb. Additional reductions in
the flexion ROM on the replaced limb contribute to the idea of increased dependence on the non-
replaced limb, which may place it at risk for further joint issues.
Pain levels were increased for the dissatisfied patients in their replaced limbs compared
to their non-replaced limbs and both other groups. It is well known that higher pain has been
associated with lower satisfaction levels (222), which is in agreement with the results of the
dissatisfied group in this study. The increased pain levels may have contributed to the
quadriceps avoidance pattern, which led to a reduced gait velocity. The total level of pain and
120
function does not seem to be as important to the patient as the relative reduction in overall pain
post-surgery (18). While the pain may manifest at the surgery site (i.e. the knee) and
subsequently contribute to dissatisfaction, additional factors have been shown to contribute to
patient satisfaction, including the need for outside assistance, lack of social interaction, or
presence of comorbidities, all of which have been shown to contribute to reduced satisfaction
levels (17). Physical comorbidities were controlled for in this study, however psychological
issues were not. It may be of benefit to further examine psychological distresses experienced by
the patients and how they impact movement and pain profiles. Addressing the psychological
issues may help patients to cope with the pain experienced and subsequently the physical
limitations. Something similar to guided imagery which can reduce the pain and anxiety
experienced post-operatively (232) may help with the psychological issues experienced which
contribute to dissatisfaction.
This study is not without its limitations. First, the ages of the TKR groups and healthy
controls were different and the age range for each of the three groups was quite wide (50 to 75).
This could have potentially influenced the results, although the average age difference was 6
years between the healthy group and the two TKR groups. Additionally, a small sample size was
used for the dissatisfied group. This may have impacted the significance of the results and future
studies may benefit from increased samples of dissatisfied patients which fit within the inclusion
criteria.
Conclusion
In summary, the level ground walking kinematics and kinetics of the dissatisfied TKR
patients do differ from satisfied patients and healthy controls. Reductions in peak loading-
response knee extension moments and abduction moments were evident in the dissatisfied group.
121
Additionally, both 1st and 2nd peak VGRF values were reduced in the replaced limb of the
dissatisfied group compared to their non-replaced limb, suggesting an asymmetry in limb
loading. Future research should examine methods to correct the asymmetrical loading
experienced by the patients as a potential means to improve patient satisfaction. This may
partially involve pain management as pain was more prevalent in the dissatisfied patients than
the satisfied patients. Both rehabilitation methods and implant design improvements should be
further evaluated for means of improving patient satisfaction.
122
Chapter III Appendix: Tables and Figures
Table 1. Peak VGRF (BW) and joint moments (Nm/kg), and knee angle and ROM (°): Mean ± STD.
Dissatisfied
Replaced
Dissatisfied
Non-
Replaced
Satisfied
Replaced
Satisfied
Non-
Replaced
Healthy
Dominant
Healthy
Non-
Dominant
Interaction
p value
1st Peak VGRF#* 1.03±0.07AC 1.08±0.07C 1.08±0.06AC 1.13±0.08 1.15±0.06 1.14±0.06 0.0041
2nd Peak VGRF#* 1.01±0.05ABC 1.04±0.05 1.07±0.05 1.09±0.05 1.10±0.08 1.09±0.07 0.0440
Flexion Loading-response
ROM#*
-11.1±6.4AC -15.4±3.3 -12.7±5.0AC -17.0±5.1 -18.1±4.1 -17.8±4.2 0.0136
Loading-response Extension
Moment#*
0.42±0.18BC 0.56±0.32BC 0.55±0.20 0.70±0.24 0.74±0.22 0.74±0.22 0.1739
Push-Off Extension Mom# 0.16±0.08 0.20±0.10 0.26±0.09 0.26±0.08 0.20±0.11 0.21±0.11 0.0721
Adduction Loading-response
ROM*
2.6±1.4 2.7±1.2 2.3±2.4C 0.9±1.4C 3.3±1.9 2.2±1.6 0.3317
Loading-response Abduction
Moment
-0.42±0.17C -0.50±0.08 -0.48±0.13 -0.48±0.14 -0.55±0.13 -0.42±0.13 0.0248
Push-Off Abduction Moment -0.29±0.15 -0.35±0.10 -0.33±0.09 -0.38±0.13 -0.33±0.10 -0.32±0.12 0.1982
Internal Rotation Loading-
response ROM
9.8±6.2A 13.8±6.8 10.6±5.6 9.7±5.1 13.8±6.9 12.5±5.8 0.0060
Loading-response Internal
Rotation Moment#
0.08±0.05ABC 0.15±0.09 0.13±0.06 0.15±0.06 0.14±0.05 0.13±0.04 0.0260
Push-Off External Rotation
Moment*
-0.08±0.04C -0.08±0.04 -0.05±0.04AC -0.08±0.04C -0.13±0.04 -0.12±0.04 0.0469
A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
123
Table 2. Ankle and Hip ROM (°) and Joint Moments (Nm/kg): Mean±STD
Dissatisfied
Replaced
Dissatisfied
Non-
Replaced
Satisfied
Replaced
Satisfied
Non-
Replaced
Healthy
Dominant
Healthy Non-
Dominant
Interaction
p value
Ankle Dorsiflexion ROM 23.9±2.1AC 20.9±3.9 24.9±3.1AC 22.1±4.3C 18.6±3.9 19.2±3.8 0.0438
Loading-response
Dorsiflexion Moment#
0.28±0.10 0.25±0.07 0.34±0.10 0.31±0.08 0.32±0.09 0.31±0.07 0.8607
Push-Off Plantarflexion
Moment#*
-1.20±0.12BC -1.25±0.14BC -1.37±0.14 -1.45±0.13 -1.39±0.14 -1.42±0.11 0.6360
Hip Extension ROM -34.7±5.4 -35.5±7.9 -34.9±5.3 -35.4±4.4 -38.1±4.0 -38.0±5.0 0.9345
Loading-response Hip
Extension Moment
-0.47±0.15 -0.52±0.12 -0.67±0.19 -0.65±0.24 -0.61±0.19 -0.55±0.19 0.1635
Push-Off Hip Flexion
Moment
0.56±0.20 0.59±0.16 0.65±0.15 0.65±0.17 0.70±0.19 0.68±0.19 0.7148
A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
124
Table 3. WOMAC Scores Pain for individual tests (0-10 VAS) Dissatisfied
Replaced
Dissatisfied
Non-Replaced
Satisfied
Replaced
Satisfied Non-
Replaced
Healthy
Dominant
Healthy
Non-
Dominant
Interaction
p value
WOMAC Total#* 794.9±484.2ABC 67.2±64.5 251.2±179.2C 196.5±175.8C 29.9±73.5 18.5±40.2 <0.0001
WOMAC Physical Function#* 525.2±323.9ABC 38.3±30.3 179.2±141.3C 144.5±131.1C 16.1±39.3 12.5±26.8 <0.0001
WOMAC Stiffness#* 76.1±58.3ABC 8.2±8.6 38.1±41.8C 26.8±44.4 6.5±18.3 2.9±6.3 0.0007
WOMAC Pain#* 193.7±138.6ABC 20.7±34.0 33.9±30.1 25.1±22.8 7.3±17.3 3.1±7.6 <0.0001
Walking Replaced limb#* 1.94±1.98ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 <0.0001
Walking Non-replaced limb#* 1.78±1.77ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 <0.0001 A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
125
CHAPTER IV
STAIR AMBULATION PATTERNS AND ASYMETRICAL LOADING OF
DISSATISFIED TOTAL KNEE REPLACEMENT PATIENTS
126
Abstract
Total knee replacement (TKR) patients have shown alterations in lower extremity
biomechanics in comparison to healthy controls during stair ascent and stair descent. However,
it is unknown how dissatisfied TKR patients differ from satisfied TKR patients and healthy
controls during more difficult activities such as stair negotiation. Therefore, the purpose of this
study was to compare knee biomechanics of dissatisfied TKR patients to satisfied TKR patients
and healthy controls during stair ascent and descent. Nine dissatisfied and fifteen satisfied TKR
patients and fifteen healthy controls participated, completing stair ascent and descent trials on an
instrumented stair case. A 2 x 3 ANOVA was used to analyze biomechanical differences
between groups and limbs during both activities. Dissatisfied patients showed reduced 2nd peak
VGRF and loading-response knee extension moments in their replaced limb compared to their
non-replaced limb and to satisfied and healthy groups during stair ascent. 1st peak VGRF and
both loading-response and push-off abduction moments showed reduced values in replaced limbs
compared to non-replaced limbs for all groups. During stair descent, the dissatisfied group
showed reduced loading-response and push-off knee extension moments in their replaced limb
compared to their non-replaced limb and the healthy group. Additional reductions in the
loading-response and push-off abduction moments in the dissatisfied group compared to the
healthy group were evident. The loading-response knee extension and abduction moments were
also reduced in the dissatisfied group compared to the satisfied group. Increased pain was
present in the replaced knee of the dissatisfied patients compared to both other groups. Future
research should examine ways to improve the symmetrical loading of the dissatisfied patients
and to alleviate the pain experienced during stair negotiation as a means of improving patient
satisfaction.
127
Keywords: total knee replacement, arthroplasty, stair biomechanics, asymmetry, satisfaction
Introduction
Stair climbing is a common activity of daily living (31) and one of the most difficult tasks
for older adults (32) due to the increased muscle and joint demands (29, 30). As patients reach
end-stage knee osteoarthritis, many frequently report difficulty with climbing stairs (33). Stair
climbing is also measured by most survey tools [including the Western Ontario and McMaster
University survey (WOMAC) and the Forgotten Joint Score (FJS)] used on the total knee
replacement (TKR) population (5, 34, 35), thereby suggesting its importance in everyday life for
most people. However, stair negotiation is less commonly studied in the TKR population
compared to level ground walking.
During stair ascent and descent, knee flexion range of motion (ROM) has been frequently
shown to be reduced in TKR patients compared to healthy controls (22, 36, 37), although one
study did report no differences between the two groups (19). TKR patients have shown reduced
peak knee extension moments during stair ascent, which is often coupled with a reduction in
speed (24), although this specific study did not investigate stair descent. Another study has
shown reductions in loading-response knee extension moments during stair ascent with no
reductions in gait speed compared to healthy controls (19). Gait speed, for example, has been
shown to increase post-operatively for TKR patients compared to pre-operative levels, however,
gait speed does not always reach levels equal to those of healthy controls at one year post-
operation (15, 38). Other studies have shown no differences between the two groups though,
with TKR patients maintaining their speed from 12-18 months as far as 46 months after surgery
(22, 26). Knee adduction angles have shown no differences for TKR patients compared to
healthy controls but an increased push-off peak knee internal abduction moment (KAM) was
128
present during stair ascent (19). Other studies have shown mixed results with some showing
increased KAM values (39) and some showing values equal to (41, 42) or reduced compared to
healthy controls (40, 41). A shift in joint loading has been seen with reduced knee extension
moments sometimes being ”transferred” to KAM (19) in an effort to keep up gait speed and
compensate for a reduced capacity in loading the knee joint.
Asymmetrical gait patterns can be problematic for TKR patients as the imbalance can
cause issues for the contralateral knee. Approximately 40-50% of unilateral TKR patients have
their contralateral knees replaced within ten years (118, 233). An examination of asymmetries in
older asymptomatic older adults revealed small insignificant levels of imbalance for the 1st and
2nd peak vertical ground reaction force (VGRF) during stair ascent (4.1% and 6.8%, respectively)
and stair descent (8.5% and 6.3%, respectively), suggesting well-balanced movement between
limbs with respect to loading (234). VGRF symmetry comparisons between TKR patients and
healthy controls revealed no significant differences, although there were evident reductions in 2nd
peak VGRF levels for TKR patients during preferred walking speeds (235). However, most
comparisons on TKR symmetry has been related to ground reaction force or temporal gait
characteristics, with no studies to date on joint kinematics and kinetics. This information may
provide added insight into the mechanical deficits between limbs for TKR patients.
For the dissatisfied TKR population, deficits in their movement profiles may be further
enhanced during more difficult activities (such as stair negotiation), suggesting a need for
examination of these activities. To date, no studies have been conducted examining the three-
dimensional biomechanical profile of dissatisfied TKR patients during stair ascent and descent
activities. Therefore, the purpose of this study was to compare the knee biomechanics profiles of
stair ascent and descent for dissatisfied TKR patients to satisfied TKR patients and healthy older-
129
adult controls. It was hypothesized that dissatisfied TKR patients would exhibit decreased knee
extension moments and increased knee abduction moments in their replaced limb compared to
their non-replaced limb and compared to the satisfied TKR patients and healthy controls during
stair ascent and descent.
Materials and Methods
Participants
Participants were recruited from a local orthopaedic clinic: nine dissatisfied TKR
patients (68.0±4.2 years, 1.69±0.07m, 80.99±18.59 kg, 34.6±14.3 months since surgery) and
fifteen satisfied TKR patients (66.6±6.3 years, 1.76±0.10m, 90.19±16.98 kg, 29.3±12.8 months
since surgery). Fifteen healthy controls (60.7±9.2 years, 1.75±0.09m, 77.74±11.75 kg) were
recruited with the same exclusion criteria as the TKR groups. The inclusion criteria for TKR
patients were having a unilateral total knee replacement (conducted by a single surgeon) at least
12 months but less than 60 months prior to testing and between the ages of 50 and 75 at the time
of testing. Potential participants were excluded if they had any additional lower extremity joint
replacements, any additional diagnosed osteoarthritis of the hip, knee, or ankle, BMI greater than
38, or neurological diseases. TKR patients were then asked (with a 5-point Likert scale available
for response), “How satisfied are you with your total knee replacement?” “Neutral” responses
were excluded. Any participant who responded “very dissatisfied” or “dissatisfied” was placed
into the dissatisfied group and participants who answered “satisfied” or “very satisfied” were
placed into the satisfied group.
Instrumentation
A twelve-camera motion analysis system (240 Hz, Vicon Motion Analysis Inc., Oxford,
UK) was used to obtain three-dimensional (3D) kinematics during all testing sessions. All
130
participants wore standardized running shoes (Noveto, Adidas, Herzogenaurach, Germany).
Anatomical retroreflective markers were placed bilaterally on the acromion processes, iliac
crests, greater trochanters, medial and lateral femoral epicondyles, medial and lateral malleoli, 1st
and 5th metatarsal heads, and 2nd toe. Four retroreflective markers clustered on a semi-rigid
thermoplastic shell was placed on the lateral aspects of both shanks and thighs, and on the
posterior aspect of the pelvis and the thoracic cage, attached via an elastic neoprene wrap. A
cluster of three retroreflective markers was placed on the dorsal aspect of both shoes and affixed
by duct tape. Anatomical and tracking markers were kept on for the static trials and anatomical
markers were removed prior to testing trials. An instrumented 3-step stair case (FP-Stairs,
American Mechanical Technology Inc., Watertown, MA, USA) was used in conjunction with
two force platforms (1200 Hz, BP600600 and OR-6-7, American Mechanical Technology Inc.,
Watertown, MA, USA) to measure the ground reaction force (GRF) and moments of forces (19).
The FP-Stairs were independently bolted to the two force platforms. An additional two non-
instrumented customized wooden steps and a landing platform were used to ensure continuous
motion after the instrumented steps (Figure 1). Each step had a depth of 29.9 cm, a width of 60.0
cm, and a rise of 17.8 cm. A handrail was available in case of loss of balance (right side during
ascent and left side during descent). Walking speed during all trials was monitored by two sets
of photo cells (63501 IR, Lafayette Instrument Inc., IN, USA) placed on the 1st and 4th steps and
electronic timers (54035A, Lafayette Instrument Inc., IN, USA).
Experimental Procedures
On day one, all participants signed an informed consent form, completed a physical
activity readiness survey (PAR-Q) to assess cardiovascular risks to exercise, and completed a
WOMAC for both knees (178)and TKR participants also completed an FJS (158). Following
131
completion of the surveys, all participants completed a five-minute walking warm up on a
treadmill at a self-selected speed. Participants were then fitted with the retroreflective marker set
described earlier. Participants then performed three practice stair ascent and descent trials at a
self-selected speed. The practice trials were to familiarize themselves with the stair set up and to
collect average ascent and descent speeds. Participants took three steps over level ground,
allowing achievement of a consistent level walking speed prior to stair ascent. A speed range
(mean ascent/descent speed ± 5%) was used to control participant speed during experimental
trials. All subjects then performed five trials in four different conditions: with the replaced limb
(TKR) or right limb (healthy) on second step (step of interest) during stair ascent, with the non-
replaced limb (TKR) or left limb (healthy) on second step during ascent, with the replaced limb
(TKR) or right limb (healthy) on second step during descent, and with the non-replaced limb
(TKR) or left limb (healthy) on second step during descent. The testing order of the limbs were
randomized, although a descent condition always followed an ascent condition in order to
minimize the number of required trials for the TKR participants. Subjects were asked to perform
the stairs in a step over manner. A trial was considered successful if the predetermined speed
range was met, and a step-over manner gait was utilized. A handrail was provided for balance
purposes if needed, but not for propulsion. Unsuccessful trials were repeated. On day two,
participants completed a five-minute warm up walking on a treadmill at a self-selected speed.
Two trials of a stair ascent/descent test using an 11-step stair case and a chair rise test were
performed, with the best times being reported. A numerical visual analogue pain (VAS) scale
was used to assess pain in both knees for TKR subjects and healthy subjects at the end of the test
conditions.
132
Data Analyses
Visual3D biomechanical analysis software suite (version 5.0, C-Motion, Inc.,
Germantown, MD, USA) was used for 3D kinematic and kinetic variable computations for stair
ascent/descent data. A Cardan rotational sequence (X-y-z) was used for 3D angular kinematics
computations and the conventions of the angular kinematic and kinetic variables were defined
using the right-hand rule. Positive values indicate ankle dorsiflexion, inversion, and internal
rotation, knee extension, adduction, and internal rotation, and hip flexion, adduction, and internal
rotation angle, ROM and internal joint moments. Kinematic and GRF data were filtered using a
fourth-order low-pass Butterworth filter with a cut-off frequency of 8Hz before kinematics and
joint moment calculations. Raw GRF were filtered separately using a fourth-order low-pass
Butterworth filter with a cut-off frequency of 50 Hz to calculate GRF variables. Critical events
and values, including peak VGRF, knee extension, abduction, and internal rotation moments,
loading-response knee flexion, adduction, and external rotation ROM, and sagittal plane hip and
ankle ROM and moments, were chosen using customized computer programs (VB_V3D and
VB_Table, MS Visual Basic 6.0, USA). Joint moments represent internally applied moments,
were reported in the proximal reference system and were normalized to body mass (Nm/kg).
GRF variables were normalized to body weight (BW). Averages across the five trials of selected
variables for every condition for each participant were used in statistical analyses.
Asymmetry indices (AI) were calculated to quantify asymmetries for peak VGRFs and
knee extension moments using the equation (236).
𝐴𝐼 =𝑋𝑅−𝑋𝑁
1
2(𝑋𝑅−𝑋𝑁)
× 100%
where XR is the replaced limb of the TKR patients or the dominant limb of the healthy
subjects and XN is the non-replaced limb or non-dominant limb of the healthy controls. An AI
133
value of zero indicates perfect symmetry, negative values indicate non-replaced limb dominance,
and positive values indicate replaced limb dominance.
Statistical Analyses
A 2x3 (limb x group) mixed model analysis of variance (ANOVA, p<0.05) was
performed to detect differences between limbs and groups for kinematic and kinetic variables,
using SAS 9.4 (Cary, NC, USA). A one-way ANOVA (p<0.05) was performed on AI,
demographic, survey, and functional test data to test for differences between the three groups.
When the ANOVA revealed a significant interaction or main effect, post-hoc comparisons with
Bonferroni adjustments were evaluated using adjusted p values of 0.00625 for 2x3 ANOVA
comparisons and 0.0167 for 1x3 ANOVA comparisons. Post hoc comparisons were only made
between the TKR replaced limbs and healthy dominant limbs and TKR non-replaced limbs
against the healthy non-dominant limbs. This was performed under the assumption that any
differences between the dominant and non-dominant limb are random and small when present in
the healthy population.
Results
During both stair ascent and descent trials, the dissatisfied group walked significantly
slower than both the satisfied and healthy groups (p=0.0006 and p=0.0005, respectively) at
preferred gait speeds. In the stair ascent condition, the dissatisfied group (0.53±0.16 m/s) was
approximately 40% slower than the satisfied group (0.74±0.10 m/s) and 30% slower than the
healthy group (0.69±0.12 m/s, p<0.0099 for both tests). In the descent condition, the dissatisfied
group (0.45±0.17 m/s) was approximately 40% slower than both the satisfied (0.64±0.08 m/s)
and healthy (0.63±0.09 m/s, p<0.0008 for both tests) groups. Additionally, during the stair
134
ascent and descent functional tests, the dissatisfied group was significantly slower than the
satisfied and healthy groups (p<0.0117 for both tests).
Significant interactions were present during stair ascent trials for 2nd peak VGRF, peak
loading-response knee extension moment, knee abduction ROM, and peak loading-response
internal rotation moment (Table 4). Post-hoc analysis revealed that the dissatisfied group had
lower 2nd peak VGRFs compared to their non-replaced limb and both the satisfied and healthy
groups (p<0.0040 for all tests). The peak loading-response extension moment was lower in the
replaced-limb of the dissatisfied group compared to their non-replaced limb and both the
satisfied and healthy groups (p<0.0041 for all tests). For all four significant interactions, the
satisfied group was different in their replaced limb compared to their non-replaced limb,
experiencing reduced VGRF, peak loading-response extension moment, and peak loading-
response internal rotation moments along with an increased knee abduction ROM (p<0.0038 for
all tests). In addition, significant limb main effects were present in the 1st peak VGRF, peak
loading-response abduction moment, and peak push-off abduction moment, with the replaced
limbs showing reduced values compared to the non-replaced limbs for all three variables (Table
4).
During stair descent, significant interactions were present for the peak loading-response
and push-off knee extension moment, peak abduction moment, and knee internal rotation ROM
(Table 5). The peak loading-response extension moment was reduced for the replaced-limb of
the dissatisfied group compared to their non-replaced limb and to both the satisfied and healthy
groups (p<0.006 for all tests). The peak push-off extension moment was reduced in the replaced
limb of the dissatisfied group compared to their non-replaced limb and the healthy group
(p<0.0032 for both tests), but not the satisfied group. The satisfied group showed a reduced peak
135
push-off extension moment in their replaced limb compared to their non-replaced limb. The
peak loading-response abduction moment was lower in the dissatisfied replaced limb compared
to the satisfied and healthy groups (p<0.0048). The peak push-off abduction moments were
lower in the replaced limb of the dissatisfied group compared to the healthy group and higher in
the non-replaced limb compared to the healthy group (p<0.00111 for both). Internal rotation
ROM was reduced in the replaced limb of the dissatisfied group compared to their non-replaced
limb (p=0.0010). A significant group main effect was present for 1st peak VGRF, with the
dissatisfied group showing reduced VGRF compared to the satisfied group, with neither group
being different from the healthy group. Significant limb main effects were present with the 2nd
peak VGRF and internal rotation moments were reduced in the replaced limbs compared to the
non-replaced limbs while the adduction ROM was increased in the replaced limbs compared to
the non-replaced limbs.
During stair ascent, AIs revealed a significantly increased 2nd peak VGRF for the
dissatisfied patients compared to healthy controls (Table 6). The AI in the loading-response
extension moment during ascent was increased in the dissatisfied group compared to both the
satisfied and healthy groups. During stair descent, the AI was increased in the dissatisfied group
for both the loading response and push-off knee extension moments compared to the healthy
group.
Discussion
The purpose of this study was to examine the knee biomechanics profiles of dissatisfied
and satisfied TKR patients in comparison to healthy controls. It was hypothesized that
dissatisfied patients would exhibit decreased knee extension moments and increased knee
abduction moments in their replaced limb compared to their non-replaced limb and both the
136
satisfied and healthy groups. The dissatisfied group did show reduced peak loading-response
knee extension moments during both stair ascent and descent in their replaced limb compared to
their non-replaced limb and both the satisfied and healthy groups, partially confirming our
hypothesis about the knee extension moments. During stair descent, the peak push-off knee
extension moment was reduced in the replaced limb of the dissatisfied group compared to their
non-replaced limb and the healthy group, also providing partial confirmation of our hypothesis
for knee extension moments.
Stair climbing has been shown to have increased demands on the involved muscles and
joints compared to activities such as walking (29), thereby highlighting the increased difficulty
of the task. External knee flexion moments have been reported as greater during stair ascent than
walking (131), indicating the increased knee joint demands. Some studies have reported reduced
knee extension moments during stair ascent for TKR patients compared to healthy controls (24,
39-41), which is in agreement with our results. However, some studies have reported no
differences in knee extension moments (22, 36, 42), although it should be noted that these studies
did not report gait speed. During similar gait speeds, one study showed no difference in either
loading-response or push-off knee extension moments (127), while an additional study showed
reduced loading-response knee extension moments in the replaced limb compared to the non-
replaced but no push-off differences during stair ascent (19). Reduced knee extension moments
have been often related to a quadriceps avoidance gait due to weaker quadriceps muscles and a
desire to avoid loading the knee joint (24, 129). The dissatisfied patients showed reduced peak
loading-response knee extension moments during stair ascent (over 30% reduction compared to
satisfied and healthy groups) coupled with a reduced gait speed of over 30% in comparison to the
137
satisfied and healthy groups. This may in part be due to reduced quadriceps strength, which has
been shown to be reduced in the dissatisfied group (Chapter V).
Similar discrepancies have been reported during stair descent with one study showing
reduced knee extension moments in TKR patients compared to healthy controls (22) while other
studies have reported no differences (36, 40, 127). Reductions in knee extension moments have
been shown to occur during simultaneous reductions in gait speed for TKR patients (24, 39-41),
which may help to partially explain the reductions in the dissatisfied patients. Dissatisfied
patients walked over 30% slower than both the satisfied and healthy groups during stair descent
and had peak loading-response knee extension moment reductions of over 40%. Past research
has shown that speed is not always the cause of reduction in knee extension moments as
reductions in peak knee extension moments have been evident compared to the non-replaced
limb with no speed differences compared to healthy controls (19), suggesting increased
compensation from the non-replaced limb. This was not the case for the dissatisfied group as the
increases in their non-replaced limb peak extension moments during ascent and descent (which
were not different from the other two groups) were not big enough to accommodate an increased
speed. Some research has shown similar speeds in TKR patients compared to healthy controls
(19) while others have reported reductions in speed for TKR patients during stair ascent (24, 39-
41). In the current study, satisfied TKR patients had similar gait speeds during stair ascent and
descent compared to healthy controls, but the dissatisfied group had reduced gait speed in both
conditions. However, pain levels were significantly increased in the replaced limb of the
dissatisfied group compared to their non-replaced limb and both the satisfied and healthy groups.
The increased pain may have been a barrier to increasing their replaced limb knee extension
138
moments and therefore their gait speed, despite higher moments in their non-replaced limbs to
compensate for the replaced limb in generation of desired speed.
Asymmetries may offer an additional insight in examining the knee biomechanics of
unilateral TKR patients. Loading-response knee extension moment AIs were increased in the
dissatisfied group (-48%) during stair ascent compared to the satisfied and healthy groups (-14%
and 3%, respectively). During stair descent, both loading-response (-55%) and push-off (-33%)
knee extension moment AI values were increased compared to healthy controls (16% and 3%,
respectively). The contralateral knee has been reported as the most common joint for
replacement after a primary TKR (118, 233, 237), with 40-50% needing the contralateral
replacement within ten years of the primary TKR (118, 233)(118, 233). Asymmetrical loading
in the joint can contribute to this as some patients will have increasing dependence on their non-
replaced limb. Sagittal plane knee moments have been shown as lower in the replaced limb
compared to the non-replaced limb in measures of asymmetry (238), which is evident in both
TKR groups during stair ascent and descent in this research. Of more significance though is the
increased levels of asymmetry in the dissatisfied group. During stair ascent, the loading-
response extension moment AI was significantly higher in the dissatisfied group (-48%)
compared to both the satisfied (-14%) and healthy (3%) groups. During stair descent, both the
loading-response and push-off knee extension moments AI were also higher for the dissatisfied
group (-55% and -33%, respectively) compared to the healthy group (16% and 3%, respectively).
The standard deviations of these AIs were quite high for the TKR groups, suggesting large
variability. This is something that should be further examined in patients during their recovery
process in order to minimize the risk of future contralateral TKR surgery. Peak VGRF AI values
were not different between the three groups, except for the 2nd peak VGRF, whereby the
139
dissatisfied group was higher than the healthy group, although the differences were much smaller
than the peak knee extension moment AI values. This is in agreement with previous research
which has shown 1st and 2nd peak VGRF asymmetries of 4.1% and 6.8%, respectively, during
stair ascent and 8.5% and 6.3%, respectively during stair descent has been shown in
asymptomatic older adults (234), suggesting a fairly equal limb balance. These values were
higher than those reported for healthy controls in this study, but similar to those reported by the
TKR groups. This lack of difference is in agreement with other research which has shown no
difference in AI between TKR patients and controls during preferred walking speed (235). So
while force loading may be the relatively symmetrical, how the joints are loaded still appears
different given the different joint moment SI values. This is further evidence for monitoring the
symmetry of patients during their recovery from a TKR procedure. In addition to the reduced
speed and given relatively equal overall loading but reduced knee extension moment symmetry,
this likely means the forces are being placed onto other joints to compensate for the replaced
knee, which may be examined in future research.
The changes in joint moments were accompanied with some changes in the VGRF levels.
First peak VGRF levels have been shown to be higher during stair descent than stair ascent (117,
119) and 2nd peak VGRF higher during ascent than descent (117), which is consistent with the
results here. During stair ascent, there were no differences in 1st peak VGRF, but decreases in
the 2nd peak VGRF in the replaced limb of the dissatisfied group compared to their non-replaced
limb and the satisfied and healthy groups. A group main effect was present for the 1st peak
VGRF during stair descent, with the dissatisfied group showing lower VGRF values than the
satisfied group, which also contributes to a reduced gait speed. What is of further interest is that
knee extension ROM during stair ascent has been shown as reduced in TKR patients compared
140
to healthy controls in multiple studies (22, 36, 37, 39-41) as has knee flexion ROM during stair
descent (22, 36), however this was not the case for either TKR group during stair ascent or
descent, as no ROM differences were evident. ROM limitations during more advanced tasks
such as stair climbing do not seem to be an issue for the dissatisfied population and this further
suggests that the pain they experience is likely not related to the motion of the knee, but the
loading of the joint experienced through the joint moments. The lack of ROM differences is in
agreement with other studies which have shown no difference when comparing TKR patients to
healthy controls during stair ascent (22) and comparing replaced to non-replaced limbs (37).
During stair descent, no differences have also been reported comparing TKR patients to healthy
controls (40) and their non-replaced limbs (37). However, there seem to be more studies with
reported differences, suggesting further examination is needed to determine what causes the
ROM differences.
Peak loading-response abduction moments were not different between groups during stair
ascent while peak loading-response abduction moments during descent were reduced in the
replaced limb of the dissatisfied compared to the satisfied and healthy group. Peak push-off
abduction moments were reduced in the replaced limb of the dissatisfied group compared to the
healthy group, providing partial confirmation of our hypotheses with respect to frontal plane
joint moments. Past research has been mixed with respect to frontal plane knee joint moments,
as it has been suggested that increased loading in the frontal plane occurs in order to increase
stability in the frontal plane of the replaced knee (19). Results with respect to this are mixed, as
some research has shown similar loading-response abduction moment results between replaced
limbs and healthy controls during stair ascent (19, 40, 41) but push-off moments have been
shown to increase in the replaced limb (19) by some research and reduced or no difference in
141
other research (28, 40, 41). These results are in partial agreement with this research as there
were no differences in either loading-response or push-off abduction moments in the dissatisfied
group compared to either the satisfied or healthy group (although a significant limb interaction
existed showing reduced loads in the replaced limbs compared to the non-replaced limbs).
Frontal plane joint moments have been suggested to play an important role in stability and
propulsion during stair ascent (19, 29, 131). In this instance, stability may be a factor as the
dissatisfied group did not have different abduction moments from the other two groups but still
showed reduced speed, suggesting the moments were employed for stability and not propulsion.
During stair descent, the internal abduction and external adduction moments have been
reported as no different in TKR patients compared to healthy controls (40, 42), which is not in
complete agreement with this study. Both loading-response and push-off abduction moments
were reduced in the dissatisfied group compared to the healthy group. Additionally, the replaced
limb of the dissatisfied group was lower than the satisfied group. Knee abduction moments have
been shown to be similar or reduced to healthy controls whether a posterior stabilized or mobile
bearing design is used (40-42, 54), so it is unlikely that implant design is a factor. In this study,
of the nine dissatisfied patients, seven had cruciate retaining (CR) designs and three had
posterior stabilized (PS) designs. Of the 15 satisfied patients, ten had CR designs, three had PS
designs, and two had bi-cruciate stabilized designs. The more likely scenario is the reduced gait
speed during stair ascent and increased pain levels forced the dissatisfied group to adjust their
gait patterns in order to try to alleviate pain and maintain their stability during a much more
difficult task such as stair descent.
In addition, the objective results presented in this study were also supported by the
subjective WOMAC and FJS scores for the dissatisfied group, which were worse compared to
142
the satisfied group. The dissatisfied group reported increased scores for all WOMAC subscales
and the total score in their replaced limb, indicating perceived reduced functional ability,
increased joint stiffness, and increased pain (see Chapter V). Their joint awareness, as evidenced
by the FJS, was significantly higher than the satisfied group, which may indicate an increased
discomfort and reduced ability of the joint to perform normally, as the dissatisfied patients are
more frequently aware of its presence. The increased WOMAC scores are in agreement with
previous literature (9, 194), with this research indicating that not only are the dissatisfied patients
subjectively worse, but also have physical moment alterations.
In this study, there were three instrumented steps with which analysis could have been
performed. The second step was chosen as the step of interest as previous research has shown
that the 1st step is different in terms of sagittal plane kinematics and kinetics in comparison to the
2nd and 3rd steps (120). The first step has been suggested to be a transition step from the level
ground walking (239) which occurs prior to the stair case and as such should not be compared
with the 2nd and 3rd steps, which have been shown to be similar (120). Additionally the size of
the steps could have an impact on the biomechanical factors, but the dimensions of the steps in
this study are similar to those of past research (126). Therefore we employed the second step as
our step of interest because it should be similar to the 3rd step.
This study has some limitations. First, stair climbing can be a difficult activity and as
such, we sought to make it as safe for our dissatisfied patients as possible. We did permit the use
of the handrail on our stair case for balance purposes, which may have affected the kinematics
and kinetics of some of the subjects in this study (two of the nine dissatisfied patients and one of
15 satisfied patients employed the handrail, and they were adamant they did not use it for
anything other than reassurance). However, we did not allow the use of the handrail for
143
propulsion purposes. Additionally, a step over manner was employed, which may be different
than some of these patients are used to when performing stair ascent and descent activities.
Conclusion
In summary, dissatisfied TKR patients do show different knee biomechanics than
satisfied TKR patients and healthy controls as well as between their replaced and non-replaced
limbs. Reduced gait speed, increased pain levels, and asymmetrical joint loading were evident in
the dissatisfied group. Future research should examine ways to improve the loading symmetry
and to alleviate the pain levels experienced by the dissatisfied patients in stair climbing related
tasks. This may help to improve the ability of the dissatisfied patients to handle increased knee
joint loads and restore functional abilities.
144
Chapter IV Appendix: Tables and Figures
Table 4. GRF and Knee Kinematics/Kinetics during Ascent
Dissatisfied
Replaced
Dissatisfied
Non-
Replaced
Satisfied
Replaced
Satisfied
Non-
Replaced
Healthy
Dominant
Healthy Non-
Dominant
Interaction
p value
1st Peak VGRF# 0.93±0.07 1.01±0.04 0.99±0.04 1.01±0.07 1.02±0.11 1.04±0.08 0.3075
2nd Peak VGRF#* 1.03±0.09ABC 1.12±0.09C 1.16±0.07A 1.22±0.11C 1.14±0.11 1.14±0.11 0.0338
Extension ROM 53.4±4.9 57.0±5.4 55.3±5.3 56.3±4.5 57.3±4.3 57.1±4.8 0.3572
Loading-response
Extension Moment#*
0.75±0.31ABC 1.11±0.26 1.08±0.26A 1.23±0.23 1.23±0.24 1.19±0.27 0.0003
Abduction ROM# -11.8±7.4 -12.2±6.7 -15.9±4.9A -9.3±5.3 -13.9±9.1 -12.5±8.0 0.0296
Loading-response
Abduction Moment#
-0.36±0.16 -0.48±0.14 -0.35±0.15 -0.40±0.17 -0.39±0.20 -0.50±0.15 0.6838
Push-off Abduction
Moment#
-0.33±0.20 -0.35±0.20 -0.27±0.18 -0.44±0.24 -0.18±0.13 -0.44±0.16 0.1561
Internal Rotation
ROM
6.2±5.2 5.4±4.4 2.3±4.2 4.7±3.9 2.0±3.5 3.1±4.8 0.2916
Loading-response
Internal Rotation
Moment#
0.23±0.16 0.30±0.22C 0.24±0.17A 0.38±0.18 0.13±0.10 0.46±0.13 0.0176
A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
145
Table 5. GRF and Knee Kinematics/Kinetics during Descent
Dissatisfied
Replaced
Dissatisfied
Non-
Replaced
Satisfied
Replaced
Satisfied
Non-
Replaced
Healthy
Dominant
Healthy Non-
Dominant
Interaction
p value
1st Peak VGRF* 1.24±0.25B 1.34±0.19B 1.52±0.19 1.59±0.19 1.44±0.09 1.43±0.09 0.2213
2nd Peak VGRF# 0.91±0.05 0.97±0.09 0.87±0.05 0.93±0.07 0.89±0.10 0.88±0.09 0.0600
Flexion ROM -82.5±3.7 -83.2±3.3 -78.8±5.7 -80.5±5.9 -80.8±5.6 -79.8±5.6 0.1576
Loading-response
Extension Moment*
0.39±0.22ABC 0.69±0.34 0.74±0.23 0.83±0.29 0.85±0.31 0.72±0.26 0.0079
Push-off Extension
Moment#
0.78±0.17AC 1.09±0.27 0.88±0.26A 1.05±0.23 0.96±0.18 0.93±0.16 0.0022
Adduction ROM# 11.2±6.9 8.6±3.4 10.4±5.3 6.8±2.8 8.5±3.4 7.6±2.5 0.4620
Loading-response
Abduction Moment
-0.40±0.17BC -0.56±0.13 -0.61±0.21 -0.55±0.23 -0.70±0.19 -0.37±0.14 0.0002
Push-off Abduction
Moment
-0.35±0.19C -0.48±0.18C -0.41±0.18 -0.36±0.14 -0.51±0.11 -0.26±0.11 0.0002
Internal Rotation
ROM#*
9.5±2.1AB 13.1±2.5BC 6.6±2.9 8.2±3.8 8.3±3.5 8.2±2.8 0.0186
Loading-Response
Internal Rotation
Moment#
0.10±0.04 0.13±0.10 0.13±0.06 0.17±0.08 0.08±0.05 0.15±0.05 0.4313
Push-off Internal
Rotation Moment#
0.14±0.07 0.20±0.10 0.18±0.08 0.20±0.07 0.12±0.05 0.20±0.06 0.2498
A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly
different from same leg of healthy group, #Limb main effect, *Group main effect.
146
Table 6. Symmetry Index for knee kinematics and kinetics during stair ascent
Dissatisfied Satisfied Healthy p-value
1st Peak VGRF (Ascent) -8.74±6.66 -2.39±8.66 -2.74±13.56 0.3132
2nd Peak VGRF (Ascent) -8.27±5.82B -5.37±8.30 -0.24±6.51 0.0274
Loading-Response Extension Moment (Ascent) -48.13±43.68AB -14.10±23.86 3.74±13.74 0.0002
1st Peak VGRF (Descent) -8.61±14.91 -4.95±13.92 0.96±7.03 0.1554
2nd Peak VGRF (Descent) -5.67±10.29 -6.45±7.88 0.19±7.52 0.0816
Loading-Response Extension Moment (Descent) -55.24±39.16B -9.35±58.74 16.35±24.79 0.0029
Push-off Extension Moment (Descent) -33.13±22.20B -17.75±32.03 2.73±14.52 0.0036 A Different from satisfied group, B Different from healthy group
147
Figure 1. The staircase with the instrumented steps (steps 1-3) and non-instrumented steps (steps 4-5) for continuous motion with a
landing platform.
148
CHAPTER V
STRENGTH AND BALANCE DEFICITS AFFECTING PATIENT SATISFACTION
FOR TOTAL KNEE REPLACEMENTS
149
Abstract
Knee strength is a common part of the total knee replacement (TKR) rehabilitation
process as an increase in knee strength is crucial for an increase in functional abilities. In
addition to strength, balance is an additional measure of TKR operation success. However, it is
unknown how dissatisfied TKR patients differ from satisfied TKR patients with respect to
strength and balance abilities. Therefore, the purpose of this study was to examine how knee
flexor and extensor strength, balance abilities, and deep knee flexion biomechanics differ for
dissatisfied TKR patients compared to satisfied patients and healthy controls. Nine dissatisfied
and fifteen satisfied TKR patients and fifteen healthy controls participated in this study,
performing isokinetic knee flexion and extension tests at 60°/s and 180°/s, bilateral and unilateral
static and dynamic balance trials, bilateral deep knee flexions, stair ascend/descend tests, and a
chair rise test. Dissatisfied patients showed reduced peak extension (180°/s) and flexion (60°/s)
torque compared to satisfied patients. No balance differences were evident, although an
increased percentage of dissatisfied patients were unable to complete the static and dynamic
unilateral balance tests. Reduced knee extension moments were evident in the replaced limb of
the dissatisfied group compared to their non-replaced limb and both satisfied and healthy groups.
Stair ascent and descent times and pain levels were increased in the dissatisfied group compared
to both other groups. An increase in strength may provide a more symmetrical movement
pattern and therefore better function. Future research should further examine the improvements
in strength and its subsequent effect on function and improving patient satisfaction.
Keywords: total knee replacement, arthroplasty, isokinetic strength, balance, deep knee flexion
150
Introduction
Current research on the dissatisfied patients who undergo a total knee replacement (TKR)
has largely focused on survey data and limited on testing of physical functions and capacities
(16-18). Survey data has shown increased difficulty in certain activities for dissatisfied TKR
populations (223) but has failed to provide additional insight into the magnitude as well as the
mechanisms of why these activities are more difficult for the dissatisfied population. For
example, stair climbing has been suggested as a difficult activity for patients with knee problems
(33), however the physical mechanisms causing difficulty are unknown. Increased strength and
balance abilities are needed in order to successfully perform more demanding daily activities
such as navigating stairs and therefore information about strength and balance for dissatisfied
TKR populations may provide insight into the difficulties experienced by dissatisfied TKR
patients. Some more challenging daily activities require higher levels of knee flexion and
increased internal knee extension moments, which is often difficult for TKR patients (11, 230).
An increase in in knee strength levels following a TKR operation is crucial for a return to
more normal function levels. Quadriceps strength is the most commonly assessed strength
variable for TKR patients: however, the hamstrings are also important for stabilizing the knee,
warranting their examination as well. Significant reductions in quadriceps and hamstrings
strength are evident early (one-month post-operatively) in the rehabilitation process, with
upwards of 60% deficits observed compared to pre-operative levels (44). By six months post-
operation, both muscle groups show significant increases in strength (45). However, TKR
patients do not normally achieve strength levels equal to those of healthy controls, with deficits
still present 12 months after surgery in the replaced limb but no difference in the non-replaced
limb compared to healthy controls (46). While most studies examine strength levels within one
151
year post-operatively, one study has shown reduced peak quadriceps and hamstrings torques at
180°/s at an average of 98 months post-operative (36), suggesting that reduced strength still
persists over time. Multiple studies involving isokinetic strength testing for this patient group
has commonly employed testing speeds of 60°/s and 180°/s (36, 142, 146). There is a lack of
research on strength with respect to TKR patient dissatisfaction. Strength deficits post-
operatively may be more pronounced in dissatisfied patients, which may impair functional
ability.
Balance is an additional measure of success for TKR operations and return to normal
daily activities as falls can be detrimental to the TKR patients (47). TKR patients have been
shown to have decreased stability after surgery compared to healthy controls (48).
Improvements in balance have been associated with improvements in functional capacities
commonly tested in stair climb, 30 second chair rise, timed-up-and-go, and improvements in gait
speed (49). Strength and balance have been often measured together as a means of using
strength gains to explain changes in balance abilities. Increased knee extensor strength coupled
with an increased gait speed has led to increased anteroposterior (AP) balance (measured through
the range of the AP trajectory of center of pressure (COP)), but increased knee extensor strength
with a reduced gait speed led to a reduced AP balance (50). However, it was shown that peak
torque did not predict single leg static balance performance (51). As with strength data, there is
a lack of balance data with respect to patient dissatisfaction.
To our knowledge, no studies have been conducted on the strength and balance abilities
as it relates to dissatisfied TKR patients. Therefore, the purpose of this study was to examine
how the knee concentric muscle strength, balance abilities, and deep knee flexion of both the
replaced and non-replaced limbs for dissatisfied TKR patients compared to satisfied TKR
152
patients and healthy controls. It was hypothesized that dissatisfied TKR patients would show
deficits of knee extensor and flexor strength, balance abilities, and internal knee extension
moment and ROM of deep knee flexion in their replaced limb compared to their non-replaced
limbs and compared to satisfied TKR patients and healthy controls.
Materials and Methods
Participants
Nine dissatisfied TKR participants (34.6±14.3 months from surgery), fifteen satisfied
TKR participants (29.3±12.8 months from surgery), and fifteen healthy participants participated
in this study (Table 7). TKR participants were recruited from a local orthopaedic clinic. The
inclusion criteria for TKR patients were the presence of a unilateral total knee replacement
performed by a single surgeon, at least 12 months but less than 60 months removed from
surgery, and between ages of 50 and 75. Exclusion criteria were any additional lower extremity
joint replacements, any additional diagnosed osteoarthritis of hip, knee, or ankle, BMI greater
than 38, or neurological diseases. TKR patients were asked, “How satisfied are you with your
total knee replacement?” The available responses were “very dissatisfied, dissatisfied, neutral,
satisfied, or very satisfied.” Neutral were excluded. “Very dissatisfied” or “dissatisfied”
responses were placed into the Dissatisfied group and “satisfied” or “very satisfied” were placed
into the Satisfied group. A healthy control group was included with the same exclusion criteria
as the TKR groups.
Instrumentation
A twelve-camera motion analysis system (240 Hz, Vicon Motion Analysis Inc., Oxford,
UK) and two force platforms (1200 Hz, BP600600, American Mechanical Technology Inc.,
Watertown, MA, USA) were used to obtain three-dimensional (3D) kinematics and GRFs during
153
deep knee flexion trials. All participants wore standardized running shoes (Noveto, Adidas,
Herzogenaurach, Germany). A cluster of four retroreflective markers was placed on the lateral
aspects of both shanks and thighs, and on the posterior aspect of the pelvis and the thoracic cage,
attached via an elastic neoprene wrap with hooks and loops of Velcro. Four individual tracking
markers were placed on the heel counter of the shoe. Anatomical retroreflective markers were
placed bilaterally on the acromion processes, iliac crests, greater trochanters, medial and lateral
femoral epicondyles, medial and lateral malleoli, 1st and 5th metatarsal heads, and 2nd toe.
Anatomical and tracking markers were kept on for the static trials and anatomical markers were
removed prior to testing trials.
Postural stability tests were performed using a balance system (Biodex Balance System
SD, Biodex Medical Systems, Shirley, New York), sampling at a rate of 20 Hz. The visual
feedback of center of pressure (COP) traces was provided to participants during practice and test
trials.
Concentric knee extension and flexion muscle strength tests were performed using an
isokinetic dynamometer (System 4, Biodex Medical Systems, Shirley, New York), speeds of
60°/s and 180°/s.
Experimental Procedures
At the beginning of the first testing session, all participants completed a Western Ontario
and McMaster University survey (WOMAC) for both knees (178). TKR participants then
completed a Forgotten Joint Score (FJS) (158). Following completion of all forms and surveys,
participants performed a five-minute walking warm up on a treadmill at a self-selected speed.
Passive knee range of motion (ROM) was then measured on both knees while participant was
lying supine on a treatment table. Participants were then fitted with the previously described
154
retroreflective marker set. Participants then performed five deep knee flexions in which they
were instructed to squat down as low as they comfortably could without fear of losing balance.
One leg was positioned on each force platform. No additional instruction was given with respect
to foot placement. The deep knee flexion trials were collected at the end of the first day as part
of a larger study with other motions not reported here.
On the second day, participants completed another five minute walking warm up on a
treadmill at a self-selected speed. Participants then completed two trials of a stair ascent/descent
test using an 11-step staircase and a chair rise test, with best times being reported. After the
completion of the functional tests, participants performed six test conditions of bilateral and
unilateral postural static and dynamic (at level 11 setting, with 1 being the most difficult and 12
being the least difficult) stability tests. Participant’s feet were placed according to the instruction
of the balance system. The participants performed three trials of 20 seconds per condition. One
practice trial per condition was given and a rest period of 30 seconds were given between trials.
A trial was repeated if the participant grabbed the handrail or required the investigator to catch
them from falling. All balance test conditions were tested in a randomized order. After the
completion of balance tests, participants completed isokinetic strength tests of the knee flexors
and extensors. Strength tests were performed last to minimize the effect of fatigue from
maximum effort muscle action on the other tests. Following a practice trial of two sub-
maximum and one maximum effort trials, participants performed three maximum effort trials of
knee flexion and extension, at two different speeds (60°/s and 180°/s). Knee flexion and
extension repetitions were tested in succession at each speed for each leg. A rest period of two
minutes were given between conditions. The speed conditions were randomized. A 0-10
155
numerical pain visual analog scale (VAS) was used to assess pain level after the tests in both
knees for TKR participants and healthy participants for all conditions and tests.
Data Analyses
Visual3D biomechanical analysis software suite (C-Motion, Inc., Germantown, MD,
USA) was used for 3D kinematic and kinetic variable computations for deep knee flexion data.
A Cardan rotational sequence (X-y-z) was used for 3D angular kinematics computations and the
conventions of the angular kinematic and kinetic variables were defined using the right-hand
rule. Positive values indicate knee extension, adduction, and internal rotation angles and joint
moments (computed as internal moments). Kinematic and GRF data were filtered using a fourth-
order low-pass Butterworth filter with a cut-off frequency of 8Hz before kinematics and joint
moment calculations. Raw GRF were filtered separately using a fourth-order low-pass
Butterworth filter with a cut-off frequency of 50 Hz to calculate GRF variables. Critical events
and values, including peak VGRF, knee extension, abduction, and internal rotation moments,
loading-response knee flexion, adduction, and external rotation ROM, and sagittal plane hip and
ankle ROM and moments, were chosen using customized computer programs (VB_V3D and
VB_Table, MS Visual Basic 6.0, USA). Joint moments represent internally applied moments,
were reported in the proximal reference system and were normalized to body mass (Nm/kg).
GRF variables were normalized to body weight (BW). Averages across the five trials of selected
variables for every condition for each participant were used in statistical analyses.
Overall, medial-lateral, and anteroposterior stability indices were calculated. Based on
the center of balance point defined as (0,0) of the X and Y coordinate of the system, X and Y
coordinates values were recorded at each sampling point (20 Hz). During the dynamic trials of
different levels of difficulty (level 11 was the dynamic level utilized in this testing), X and Y
156
coordinate data were scaled at 1/100th degrees, up to 20 degrees of the test platform’s tilt. The X
and Y coordinates effectively represent mediolateral (X) and anteroposterior (Y) deviations from
the center of the balance surface. The Overall Stability Index (OSI) was computed as:
𝑂𝑆𝐼 = √Σ(𝑋)2−Σ(𝑌)2
𝑁
where X is the mediolateral coordinate, Y is the anteroposterior coordinate, and N is the number
of data points sampled. The Mediolateral stability index (MLSI) was calculated through the
following equation:
𝑀𝐿𝑆𝐼 = √Σ(𝑋)2
𝑁
The anteroposterior stability index (APSI) was calculated through the following equation:
𝐴𝑃𝑆𝐼 = √Σ(𝑌)2
𝑁
A customized Matlab script (MathWorks, Natick, MA, USA) was created to calculate
strength related variables, including peak torque and loading rate. Loading rate was calculated
by dividing the peak torque by its time (from onset of the movement).
Statistical Analyses
A 2 x 3 (limb x group) mixed model analysis of variance (ANOVA, p<0.05) using SAS
(Version 9.4, Cary, NC, USA) was performed to detect differences between limbs and groups for
selected variables of deep knee flexion, balance, strength variables, and WOMAC. A one-way
ANOVA (p<0.05) was performed on demographic, survey, bilateral balance, and functional test
data to test for differences between the three groups. When the ANOVA results revealed a
significant interaction or main effect, pairwise t-tests were used to compare means with adjusted
p values of 0.00625 for 2x3 ANOVA significant interactions and 0.0167 for 1x3 ANOVA
157
comparisons. Post hoc comparisons were only made between the TKR replaced and healthy
dominant limbs and TKR non-replaced limbs against the healthy non-dominant limbs. This was
performed under the assumption that any differences between the dominant and non-dominant
limb are random and small when present in the healthy population.
Results
Post hoc tests revealed for the functional tests, the dissatisfied group had increased stair
ascent (p=0.0148) and descent (p=0.0031) times compared to both the satisfied and healthy
groups (Table 7). Significant interactions were present for pain levels during chair rise, stair
ascent, and stair descent tests ; and the dissatisfied group reported increased pain in their
replaced limb compared to other groups and respective limbs during all three tests (p<0.0148 for
all tests, Table 12). Significant interactions revealed the dissatisfied group had a decreased
passive knee ROM in the replaced limb compared to their non-replaced limb and the healthy
group (p<0.0015 for both tests) but no difference from the replaced limb of the satisfied group
(Table 8). Significant interactions were present for all WOMAC subscales and total scores
(Table 8). The dissatisfied group reported increased total WOMAC and subscale scores in their
replaced limb compared to their non-replaced limb and satisfied and healthy groups. The
satisfied group reported increased WOMAC total and physical function scores for both limbs as
well as increased stiffness in their replaced limb, compared to the healthy group.
Significant interactions were present for peak torque of isokinetic knee extension strength
at 180°/s and flexion at both 60°/s and 180°/s (Table 9). The post hoc comparisons showed that
the dissatisfied replaced limb had lower peak extension torque at 180°/s than the non-replaced
limb of the dissatisfied group, and both satisfied and healthy groups (p<0.0051 for all tests).
Peak torque during the 60°/s flexion test was reduced in the replaced-limb of the dissatisfied
158
group compared to that of the satisfied group (p=0.0168). Additionally, there was a limb main
effect for the 60°/s extension test, with the replaced limbs having lower peak torque than the
non-replaced. However, there were no differences in loading rate for any of the tested speeds.
Significant interactions revealed that the dissatisfied group reported increased knee pain in their
replaced limb during the 60 (p=0.0018) and 180°/s (p=0.0011) strength tests compared to their
non-replaced limb and satisfied and healthy groups. (Table 12). A group main effect was present
during the non-replaced 60°/s test, with the dissatisfied group showing increased pain compared
to the satisfied and healthy groups.
There were no interactions for any of the balance variables (Table 10). There was a limb
main effect for the ML Stability index during the static condition, with replaced limbs having
increased ML sway during unilateral stance. It is of interest to note, approximately 33% of
dissatisfied participants were unable to complete unilateral static balance trials on either limb
while 13% of the satisfied group were unable to do so on their replaced limb and 20% on their
non-replaced limb. During the dynamic unilateral tests, these numbers increased to 56% of
dissatisfied participants and 27% of satisfied participants on both limbs. All healthy participants
were able to complete the test on the non-dominant limb but 13% could not on their dominant
limb. The dissatisfied group reported increased knee pain in their replaced knees compared to
their non-replaced knee and satisfied and healthy groups during all balance tests (except for
unilateral balance on the non-replaced static trials, Table 12).
During active knee flexion, there were significant interactions for peak knee extension
moment, peak knee abduction moment, and knee internal rotation ROM (Table 11). The
replaced limb of the dissatisfied group showed a decreased peak knee extension moment
compared to their non-replaced limb, the replaced limb of the satisfied, and the healthy groups,
159
while the satisfied group only showed a decreased peak knee extension moment in their replaced
limb compared to the non-replaced limb. For peak knee abduction moment, post hoc analysis
revealed no significant differences. The replaced limb of the satisfied group showed a reduced
knee internal rotation ROM compared to the healthy group. There was a limb main effect for
knee adduction ROM, with the non-replaced limb of TKR groups or non-dominant limb of
healthy group being lower than the replaced or the dominant limb (Table 11). There was a group
main effect for knee pain during the active flexion, with the dissatisfied group reporting
increased pain levels in both knees compared to the other groups (Table 12).
Discussion
The purpose of this study was to examine differences in strength, balance, and deep knee
flexion between dissatisfied and satisfied TKR patients. Our hypothesis related to reductions in
knee extension strength for the replaced limbs of the dissatisfied group was partially confirmed
as the dissatisfied group showed reductions in peak knee extension torque at 180°/s of the
replaced limbs compared to their non-replaced limbs, those of the satisfied and healthy groups.
The limb main effect for knee extension during the 60°/s test showed reduced extension strength
for the replaced limbs compared to the non-replaced limbs, offering partial support for our
hypothesis. Additionally, the decreased peak flexion torque in the dissatisfied replaced limb at
60°/s compared to the satisfied group also provided partial confirmation of our hypothesis. Post-
operative increases in knee extensor and flexor strength have been related to improvements in
balance and restoration of functional abilities (240-242). It has been suggested that knee
extensor strengthening exercises should continue long-term to improve patient satisfaction and
restore functional abilities (243). Decreased concentric strength in the knee extensors has been
linked to decreased stair climbing abilities (240), which is in partial agreement with the results of
160
this study. The dissatisfied group showed reduced stair ascent and descent times compared to the
satisfied and healthy groups while having reduced knee extensor strength at 180°/s in their
replaced limb. The replaced limbs showed lower peak torque compared to the non-replaced
limbs during knee extension at 60°/s. This imbalance in the dissatisfied group (30% strength
increase in the non-replaced limb compared to the replaced limb) is substantially larger than that
in the satisfied group (12% strength increase in non-replaced limb) which could render the given
task of stair climbing more difficult (with only a 4% difference between dominant and non-
dominant limbs for the healthy group). Our results of knee extensor and flexor peak torque
differences at both 180°/s and 60°/s are in partial agreement with previous research which has
shown peak torque reductions in TKR patients compared to healthy controls for both knee
extensors and flexors at 180°/s but not 60°/s (36). While not all results were statistically
significant, it is worth noting that average peak torque values across the groups and limbs were
always lower in the dissatisfied group compared to the satisfied and healthy groups, for both the
replaced and non-replaced limbs. It can be generally concluded that reduced strength is present
in the dissatisfied TKR population.
The reductions in strength levels may be in part related to the increased pain levels
present in the replaced limb of the dissatisfied group. Pain relief and subsequent restoration of
activity abilities have been highly correlated with increasing patient satisfaction (241). Daily
activities with increased difficulty often require increased knee flexion and extension
movements, which often require increased quadriceps and hamstring strength (242). This can be
further confounded when asymmetries exist. Limb loading asymmetries have been present early
after surgery and have been shown to contribute to increased stair climbing times (244). These
reductions in functional abilities have been speculated as being related to pain and quadriceps
161
weakness, both of which are present in our dissatisfied patient population. Quadriceps weakness
is often a point of interest during the rehabilitation process, and an increase in the strength levels
with a subsequent increase in knee ROM have been evident with increased patient satisfaction,
although this same study did not show any differences in WOMAC pain and function scores
associated with the strength and ROM gains (245). It should be noted that this previous study
did not group the patients based on their satisfaction as has been done in the current study. The
current study showed reductions in passive knee ROM for the dissatisfied group in their replaced
limb compared to all other groups and respective limbs, except for the replaced limb of the
satisfied group. In addition, the increased WOMAC total and sub-scale scores showed increased
deterioration of the replaced limb of the dissatisfied group.
A goal of TKR implants is to increase maximum knee flexion of TKR patients, with the
idea that increased flexion leads to improved patient outcomes (246). During weight-bearing
flexion related activities, over 85% of TKR patients achieve flexion ROM over 100° (247) but
this was not the case in the deep knee flexion test for either TKR group in this study.
Participants in this study were instructed to squat as low as possible in the deep knee flexion test,
and neither TKR group achieved mean flexion ROM above 100°, although during this activity,
no statistically significant flexion ROM differences were present. However, asymmetries were
present in peak knee extension moments for the dissatisfied TKR group. Additionally, the
dissatisfied group had a reduced peak knee extension moment in their replaced limb compared to
the satisfied and healthy groups, likely due to the increased pain levels experienced during the
activity. What is of interest though is that they also experienced increased pain in their non-
replaced limb during the activity, likely due to the increased joint loading experienced by
compensating with their non-replaced limb. This increased pain level on both sides can make the
162
task increasingly more difficult to achieve. It may ultimately lead to an avoidance to this type of
high flexion task, further increasing their dissatisfaction with the replaced knee. It is paramount
during the rehabilitation process to promote achievement of these higher levels of deep knee
flexion in order to increase long-term outcomes. An inability to do so may be a good
identification of patients at-risk of poor long-term outcomes (248). Deep knee flexion is a
difficult task and it has been speculated that a return to normal function during difficult activities
such as deep knee flexion is not truly possible for TKR patients (249). TKR patients may be
able to achieve a shallower knee flexion compared to healthy individuals, possibly due to the
pain experienced during the activity. However, alleviating the pain may allow for more
symmetrical function and allow for task completion. It has been reported that 75-86% of TKR
patients have knee symptoms during squatting, as 42-59% have moderate to severe difficulty
during squatting, and 25% are unable to squat at all (11, 230). This is in agreement with the
current results showing the asymmetry in deep knee flexion as well as the presence of pain for
the dissatisfied group.
No significant differences were present in both the static and dynamic balance levels and
limbs, aside from slight increases in ML stability in the replaced limb compared to the non-
replaced limb. This was not in agreement with our hypothesis. Additionally, this is of interest
due to the significant number of patients who could not complete the balance trials. During the
unilateral static balance test, 33% of dissatisfied patients were unable to complete the tests on
both their replaced and non-replaced limbs, which is an increase over the 13% and 20% of the
satisfied group on their replaced and non-replaced limbs, respectively. As the difficulty
increased to the dynamic unilateral stability test, these numbers elevated to 56% of dissatisfied
and 27% of satisfied patients. This may be in part related to replacement design types received
163
by our patients in the study. Most of the participants had cruciate retaining TKR designs, which
do not contain a cam-post mechanism. Longer cam-post designs aid frontal plane stability to the
replaced knee joint (250). Without this added stability, the replaced knees are left to be
controlled by musculature and ligaments, some of which (knee musculature) has been shown as
reduced in the dissatisfied patients. Further research is warranted to examine the effects of
different TKR design types on patient balance ability. With these reduced numbers of patients
who were unable to complete the tests, our sample sizes for the statistical analysis were greatly
reduced, which may have skewed results of the balance tests as no differences were reported
when the patient was able to complete the trials. A completion rate of less than 50% for the
dissatisfied patients is quite low, although the difficulty level was set at the lowest level (level
11) in the unilateral dynamic balance test. The inability to perform the balance tests may have
also contributed to the patient dissatisfaction. Additionally, pain levels were increased for the
replaced limb of the dissatisfied participants compared to the satisfied and healthy groups.
Improvements in balance have been associated with improvements in gait speed, chair rise tests,
and stair climb tests (49). The presence of pain may not be enough to impact the balance
abilities at the difficulty levels tested here when patients were able to complete the test.
However, the subsequent improvements in stair climbing were not evident in this research.
Strength and balance are frequently connected since strength is necessary to maintain balance.
Increased knee extensor strength coupled with an increased gait speed has led to increased
anteroposterior balance, but increased knee extensor strength with a reduced gait speed led to a
reduced anteroposterior balance (50). This was not evident with the dissatisfied group as no AP
differences were evident. Other research has shown that peak torque did not predict single-limb
164
balance (51), which does agree with the results of this research since strength differences were
evident but balance differences were not.
The consistent issue with the dissatisfied patients is the presence of pain, which
subsequently seems to impact certain physical abilities. Increased patient dissatisfaction has
been associated with increased pain during activities, lower knee function post-operatively, and
depressive symptoms or somatization dysfunction (251). Treating the psychological factors
associated with the surgery may help to improve outcomes. Patients with somatization
dysfunction have a two-fold increase in risk for dissatisfaction with their TKR (251). This may
suggest one of two options. First, pre-operative psychological screening may help with the
rehabilitation of TKR patients as certain psychological needs can be addressed in conjunction
with the physical needs. Second, psychological tools to deal with pain management may help to
improve functional ability. Multivariate logistic and linear regression models have suggested
that dissatisfaction is largely associated with the pain reductions and functional improvements
experienced during the first three months post-operatively, suggesting that the decision to have
surgery should be based on functional limitations, not high pain scores (252). However, given
that pain is present and may continuously impair functional abilities, it should be addressed
during rehabilitation processes. Tools such as guided imagery may help to lower pain and
anxiety experienced post-operatively (232). A reduction in pain experienced by dissatisfied
patients may help improve their functional abilities and subsequently increase their satisfaction
levels. Given the physical and time-based investment in an operation such as a TKR, it is
beneficial to address all issues which may improve outcomes.
This study has some limitations. First, given the large amount of variables and
comparisons in our statistical analyses, type I error is a potential issue. Some of these
165
differences detected in this study may be by pure chances. Second, the classification of the
patient satisfaction and dissatisfaction is subjective in nature which is likely a multifaceted
response based on factors such as personal experience, bias, and perception, none of which were
captured in this study. It is simply a question posed to the patients about their perception of their
replaced limb without qualifying whether the satisfaction is based on pain, functional ability, or
any other characteristics. As mentioned earlier, the balance results are to be interpreted with
caution given the small sample sizes due to the inability of some dissatisfied patients to complete
the test. Finally, while patients were asked to perform isokinetic tests with maximum effort,
there may have been some reduced performances due to the presence of pain. With the presence
of pain, these results appeared to be their maximum efforts, it may not reflect the true muscular
abilities of the patients.
Conclusion
Patient dissatisfaction is a complex construct that is evidently multifaceted. Throughout
all tests, pain was evident, regardless of changes in function. Strength deficits were evident for
the dissatisfied patients in both the knee flexors and extensors. Balance abilities were also
reduced through an inability to successfully balance on one limb, although when possible, no
differences were evident between groups who were able to complete the unilateral balance tests.
Self-reported increased difficulty with certain activities for dissatisfied patients was evident
through a structural imbalance during deep knee flexion tests, whereby dissatisfied patients
exhibited an increased dependence on their non-replaced limb. An increase in strength may
provide a more symmetrical movement pattern and therefore better function. This may improve
patient satisfaction. More research needs to further determine the mechanisms contributing to
patient dissatisfaction and subsequent ways to improve them. This may be achieved through
166
longitudinal tracking of patient satisfaction in an attempt to identify both the physical and
psychological factors as they occur over time.
167
Chapter V Appendix: Tables and Figures
Table 7. Descriptive statistics, functional tests, and survey data.
Dissatisfied Satisfied Healthy p value
Age (years) 68.0±4.2A 66.6±6.3A 60.7±9.2 0.0034
Height 1.69±0.07 1.76±0.10 1.75±0.09 0.1280
Weight 80.99±18.59 90.19±16.98 77.74±11.75 0.0944
BMI (kg/m2) 28.13±4.61 28.85±4.26 25.33±3.34 0.0563
Months from surgery 34.6±14.3 29.3±12.8 NA 0.3598
FJS (replaced limb) 21.53±16.04B 67.78±27.76 NA 0.0002 Chair Rise (s) 18.43±7.26 16.84±5.45 15.01±4.21 0.3327
Stair Ascent (s) 5.50±1.93AB 4.30±0.79 4.06±0.68 0.0117
Stair Descent (s) 5.93±2.91AB 3.98±0.60 3.68±0.50 0.0021 A Different from Healthy group, B Different from Satisfied group.
Table 8. WOMAC Scores (100mm VAS) and Passive Knee ROM (°). Dissatisfied
Replaced
Dissatisfied
Non-Replaced
Satisfied
Replaced
Satisfied Non-
Replaced
Healthy
Dominant
Healthy
Non-
Dominant
Interaction
p value
WOMAC Total#* 794.9±484.2ABC 67.2±64.5 251.2±179.2C 196.5±175.8C 29.9±73.5 18.5±40.2 <0.0001
WOMAC Physical Function#* 525.2±323.9ABC 38.3±30.3 179.2±141.3C 144.5±131.1C 16.1±39.3 12.5±26.8 <0.0001
WOMAC Stiffness#* 76.1±58.3ABC 8.2±8.6 38.1±41.8C 26.8±44.4 6.5±18.3 2.9±6.3 0.0007
WOMAC Pain#* 193.7±138.6ABC 20.7±34.0 33.9±30.1 25.1±22.8 7.3±17.3 3.1±7.6 <0.0001
Passive Knee ROM# 118.3±7.4AC 131.0±10.4 122.9±10.5AC 134.4±9.3 134.3±13.5 135.4±13.4 0.0001 A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
168
Table 9. Peak Isokinetic Knee Extension and Flexion Torque (Nm) and Extension and Flexion Loading Rate (LR; Nm/s). Dissatisfied
Replaced
Dissatisfied
Non-Replaced
Satisfied
Replaced
Satisfied Non-
Replaced
Healthy
Dominant
Healthy Non-
Dominant
Interaction
p value
Extension @ 60°/s # 85.2±30.9 110.5±41.1 118.3±39.1 132.5±41.3 117.6±36.4 122.1±39.3 0.0664
Extension @ 180°/s# 58.3±22.7ABC 72.9±26.9 83.0±24.9A 95.4±27.0 80.8±29.3 76.4±25.8 0.0081
Flexion @ 60°/s 48.0±15.9B 52.4±19.8 64.3±21.1 67.1±21.0 63.2±13.0 58.0±17.1 0.0168
Flexion @ 180°/s* 34.4±13.7 33.7±15.3B 48.5±19.0A 57.0±21.4 47.7±14.8 45.8±12.7 0.0020
Extension LR @ 60°/s 151.0±114.0 139.2±68.9 169.1±95.9 189.4±73.4 164.8±76.2 158.3±61.9 0.6898
Extension LR @ 180°/s 244.2±92.3 261.8±103.8 343.5±160.5 342.1±162.6 274.9±132.1 243.3±104.6 0.7562
Flexion LR @ 60°/s 135.3±81.6 190.2±207.5 202.2±150.1 163.6±86.9 195.8±246.4 223.9±329.5 0.6871
Flexion LR @ 180°/s 185.0±84.0 171.0±111.1 288.7±184.2 320.2±224.0 295.5±213.9 224.9±203.2 0.2566
A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
Table 10. Unilateral overall stability index (OSI), mediolateral stability index (MLSI) and anteroposterior stability index (APSI) and
bilateral OSI, MLSI, and APSI stability indices (one-way ANOVA). Dissatisfied
Replaced
Dissatisfied
Non-
Replaced
Satisfied
Replaced
Satisfied Non-
Replaced
Healthy
Dominant
Healthy Non-
Dominant
Interaction
p value
Static OSI 2.68±1.17 3.20±1.35 3.05±1.59 2.21±1.46 3.08±1.52 1.71±0.81 0.1314
Dynamic OSI 2.23±0.57 2.60±0.47 2.55±0.81 2.56±0.92 2.38±0.52 2.41±0.84 0.8397
Static APSI 1.55±0.90 2.40±1.20 1.72±1.36 1.38±1.16 1.50±1.26 1.14±0.63 0.2484
Dynamic APSI 1.65±0.95 2.38±0.39 1.45±0.73 1.61±0.85 1.73±0.81 1.49±0.80 0.1316
Static MLSI# 1.95±1.04 1.80±0.95 2.20±1.20 1.39±1.14 2.35±1.28 1.03±0.55 0.2598
Dynamic MLSI 0.90±0.88 0.93±0.28 1.64±1.12 1.50±1.18 1.27±0.51 1.57±0.85 0.7045
Static OSI Bilateral 1.36±1.86 0.90±0.86 1.40±1.36 0.5498
Dynamic OSI Bilateral 1.69±0.53 1.56±0.75 1.79±0.65 0.6511
Static APSI Bilateral 1.13±1.89 0.70±0.85 1.06±1.20 0.6518
Dynamic APSI Bilateral 1.24±0.54 1.21±0.79 1.39±0.74 0.7786
Static MLSI Bilateral 0.39±0.43 0.35±0.36 0.65±0.84 0.3504
Dynamic MLSI Bilateral 0.90±0.43 0.73±0.35 0.85±0.27 0.4439 A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
169
Table 11. Deep knee flexion kinematics and kinetics. Dissatisfied
Replaced
Dissatisfied
Non-
Replaced
Satisfied
Replaced
Satisfied Non-
Replaced
Healthy
Dominant
Healthy Non-
Dominant
Interaction
p value
VGRF (BW) 0.56±0.07 0.61±0.05 0.60±0.08 0.64±0.08 0.63±0.08 0.62±0.08 0.0895
Knee Flex ROM (°) -86.2±19.6 -88.4±21.7 -94.9±13.3 -95.7±12.9 -100.5±23.4 -100.4±21.6 0.6972
Knee Add ROM (°) # 19.0±12.2 19.5±7.8 19.2±7.0 13.2±6.9 19.9±8.3 13.2±7.5 0.2252
Knee Int Rot ROM (°) 14.5±11.0 13.0±9.2 9.7±8.0C 12.0±6.9 20.3±13.2 13.1±9.5 0.0312
Knee Ext Mom (Nm/kg)* 0.75±0.14ABC 0.89±0.16 0.98±0.24A 1.09±0.23 1.05±0.24 1.00±0.29 0.0417
Knee Abd Mom (Nm/kg) -0.29±0.22 -0.32±0.12 -0.33±0.16 -0.29±0.18 -0.41±0.25 -0.29±0.15 0.0292
Knee Ext Rot Mom (Nm/kg) -0.13±0.06 -0.12±0.08 -0.18±0.08 -0.16±0.09 -0.24±0.12 -0.18±0.08 0.3769 A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
Table 12. Pain for individual tests (0-10 Likert). Dissatisfied
Replaced
Dissatisfied
Non-Replaced
Satisfied
Replaced
Satisfied Non-
Replaced
Healthy
Dominant
Healthy
Non-
Dominant
Interaction
p value
Chair Rise#* 1.44±2.00ABC 0.00±0.00 0.13±0.52 0.13±0.52 0.00±0.00 0.00±0.00 0.0013
Stair Ascend#* 1.56±1.94ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0001
Stair Descend#* 1.78±2.44ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0006
Bilateral Static Balance#* 1.33±2.06ABC 0.22±0.67 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0033
Bilateral Dynamic Balance#* 0.78±1.56ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0257
Unilateral Replaced static#* 1.56±2.46ABC 0.22±0.67 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0064
Unilateral Replaced Dynamic#* 0.78±1.56ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0257
Unilateral Non-replaced static 0.33±1.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.07±0.26 0.1508
Unilateral Non-replaced Dynamic#* 0.78±1.56ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0257
Replaced 60°/s#* 1.67±2.5ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0018
Replaced 180°/s#* 1.56±2.24ABC 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.0011
Non-replaced 60°/s * 0.78±1.56 0.11±0.33 0.00±0.00 0.00±0.00 0.00±0.00 0.10±0.39 0.0615
Non-replaced 180°/s 0.78±1.56ABC 0.00±0.00 0.00±0.00 0.07±0.26 0.00±0.00 0.07±0.26 0.0197
Deep Knee Flexion* 1.39±1.69 1.83±2.89 0.07±0.26 0.27±0.59 0.07±0.26 0.07±0.26 0.7632 A Significantly different from contralateral leg of same TKR group, B Significantly different from same leg of satisfied TKR group, C Significantly different from
same leg of healthy group, #Limb main effect, *Group main effect.
170
CHAPTER VI
LOGISTIC REGRESSION ANALYSES REGARDING PATIENT DISSATISFACTION
WITH TOTAL KNEE REPLACEMENT OUTCOMES
171
Abstract
Current research on patient satisfaction after a total knee replacement (TKR) lacks an
examination of objective assessments with respect to gait biomechanics, strength, and balance
abilities. Therefore, the purpose of this research was to examine associations between patient
satisfaction and the gait biomechanics, strength, balance, functional capacities, and survey data.
Twenty four TKR patients participated in overground walking, stair ascent and descent,
isokinetic strength, static and dynamic balance, and functional tests. Nine patients were in the
dissatisfied group and fifteen in the satisfied group. Logistic regression analyses were performed
to identify four models of satisfaction prediction: one for walking biomechanics, stair ascent
biomechanics, stair descent biomechanics, and functional/survey data. The functional model was
inclusive of WOMAC total scores, stair ascent and chair rise times, and peak knee extension
torque. The walking model included 1st and 2nd peak VGRF, knee extension moment, and the
forgotten joint score. The stair ascent model included 2nd peak VGRF, knee extension moment,
preferred gait speed, and peak extension torque. The stair descent model included knee
extension moment, preferred gait speed, peak extension torque, and the forgotten joint score.
The biomechanical models included both VGRF and knee extension moments, indicating their
relevance to patient satisfaction. Pain was not included in any models (as in previous prediction
models) due to a complete separation of data points.
Keywords: logistic regression, total knee replacement, satisfaction, arthroplasty
172
Introduction
Patient dissatisfaction is a complicated and multi-faceted construct involving both
subjective and objective information (16). Many of the subjective survey tools utilized in the
total knee replacement (TKR) population examine patients’ desires and abilities to return to
performing more advanced activities (111, 210). This suggests that there is an expectation that
the surgery will help the patient to return to these activities. The desired outcomes are not
always evident though as some patients have difficulty with function and lingering pain in the
replaced knee joint. Overall, most TKR operations are considered successful as there are often
reductions in pain levels and improvements in ROM in the replaced knee (6-9). However, patient
satisfaction rates for the procedure have been reported between 81-97% (10, 11). This leaves a
significant portion of the TKR population as being dissatisfied with the outcomes of their
replaced knees. Post-operative pain (12) and functional limitations (13) are commonly reported
by dissatisfied patients. Reduced pain and increased functional ability are often seen as a
defining point for “success” of the operation as they are deemed to determine the restoration of
function for the replaced joint. However, these test results do not sufficiently explain why the
TKR patients are dissatisfied with the TKR outcomes, thereby suggesting that additional
research into the mechanisms of dissatisfaction is needed. Current literature is lacking on the
biomechanical studies related to patient dissatisfaction and the variables which most contribute
to dissatisfaction.
Current research on the dissatisfied population has primarily focused on survey data and
minimally on biomechanical testing (16-18). Lab-based biomechanical studies focusing on level
walking, stair negotiation, strength, and balance can provide detailed and quantifiable insights
into the movement profiles of the dissatisfied TKR population. For example, one study has
173
shown no difference in functional test scores while showing reduced loading-response knee
extension moments and increased push-off knee abduction moments in the replaced limb of TKR
patients compared to the non-replaced limbs and healthy controls during stair ascent (19).
Despite similarities in functional tests, the biomechanical analysis was able to show deficits in
the movement patterns, providing valuable and detailed gait biomechanics in movement
recovery, which would not be detectable. Prior to the collection of the data set employed in this
research, there was a lack of three-dimensional kinematic and kinetic research specifically on the
dissatisfied TKR patient population. Through this data set, kinematic, kinetic, strength, and
balance deficits in the dissatisfied patient population have been identified, including reduced
knee extensor moments during level walking, stair ascent, and stair descent, reduced gait speed
in all three activities, reduced stair ascent/descent performance tests, reduced knee
extensor/flexor strength, and differences in survey data compared to the satisfied patient
population.
The research on patient satisfaction currently lacks an examination of objective and
numerical assessments with respect to gait biomechanics, strength, and balance abilities. This
presents a unique opportunity to identify the physical differences which may help to improve
TKR patient satisfaction rates. The purpose of this research was, therefore, to examine
associations between TKR patient dissatisfaction and gait biomechanics, strength, balance,
functional capacities, and survey data (measuring joint awareness, pain, stiffness, and functional
ability) using a logistic regression analyses. Four subsequent models were created to reflect data
traditionally associated with patient satisfaction (survey and functional test data) and three
additional models for overground walking, stair ascent, and stair descent data. It was
hypothesized that reduced knee extensor moments, reduced quadriceps strength, reduced gait
174
speed, and increased pain would contribute to increased dissatisfaction in TKR patients in the
biomechanical models.
Materials and Methods
Participants
There were 24 TKR participants recruited from a local orthopaedic clinic. The inclusion
criteria for TKR patients was the presence of a unilateral total knee replacement performed by a
single surgeon, at least 12 months but less than 60 months removed from surgery, and between
the ages of 50 and 75. Potential participants were excluded if they had any additional lower
extremity joint replacements, any additional diagnosed osteoarthritis of the hip, knee, or ankle,
BMI greater than 38 or neurologic diseases. TKR patients were then asked, “How satisfied are
you with your total knee replacement?” The available responses were “very dissatisfied,
dissatisfied, neutral, satisfied, or very satisfied” using a 1-5 Likert scale. Participants who
responded neutral were excluded, thus shifting the scale to a 1-4 scale for the “very dissatisfied,
dissatisfied, satisfied, or very satisfied” responses, respectively. Participants were subsequently
categorized into two groups based on their responses to the satisfaction question: “very
dissatisfied” and “dissatisfied” were grouped together as were “satisfied” and “very satisfied”
responses. Nine TKR patients (68.0±4.2 years, 1.69±0.07m, 80.99±18.59 kg, 34.6±14.3 months
post-surgery) were grouped as dissatisfied patients and fifteen TKR patients (66.6±6.3 years,
1.76±0.10m, 90.19±16.98 kg, 29.3±12.8 months post-surgery) were grouped as satisfied.
Experimental Procedures
On day one, all participants signed an informed consent form, completed a physical
activity readiness survey (PAR-Q) to assess cardiovascular risks to exercise, and completed a
Western Ontario and McMaster University survey (WOMAC) for both knees (178) and the
175
Forgotten Joint Score (FJS (158)). Following completion of the surveys, all participants
completed a five-minute warm up on a treadmill, walking at a self-selected speed. Participants
were then fitted with retroreflective discrete anatomical and clustered tracking markers affixed to
their thoracic cage, pelvis, thighs, shanks and feet. Three-dimensional kinematics and kinetics
were obtained using a twelve-camera motion analysis system (240 Hz, Vicon Motion Analysis
Inc., Oxford, UK), two force platforms (1200 Hz, BP600600 and OR-6-7, American Mechanical
Technology Inc., Watertown, MA, USA), and an instrumented 3-step stair case (FP-Stairs,
American Mechanical Technology Inc., Watertown, MA, USA). Participants performed five
trials of stair ascent and five trials of descent with their replaced limb on the second step of the
five step stair case, using a step-over manner. Following completion of stair trials, five
overground walking trials were performed, with the replaced limb coming into contact with the
force platform.
On day two, participants completed a five-minute warm up walking on a treadmill at a
self-selected speed. Passive knee range of motion (ROM) was then measured on both knees
while participant was lying supine on a table. Two trials of a stair ascent/descent test using an
11-step stair case and a chair rise test were performed, with average times being reported. After
the completion of the functional tests, participants performed bilateral and unilateral postural
stability tests in a static position and at level 11 on a Biodex Balance System SD (Biodex
medical Systems, Shirley, NY, USA). Participant’s feet were placed according to the instruction
of the balance system. The participants performed three trials per condition, with each trial
lasting 20 seconds. If the participant was unable to complete a condition, no score was given.
All balance test conditions were tested in a randomized order. After the completion of balance
tests, participants completed isokinetic strength tests of the knee flexors and extensors. Strength
176
tests were performed last to minimize the effect of fatigue. Following a practice trial of two sub-
maximum and one maximum effort practice trials, participants performed three maximum effort
trials at two different speeds (60°/s and 180°/s), testing the concentric strength of the knee
flexors and extensors. A rest period of two minutes was given between conditions. The speed
conditions were randomized. A visual analogue pain scale was used to assess pain in both knees
for all conditions and tests.
Data Analyses
Visual3D biomechanical analysis software suite (version 5.0, C-Motion, Inc.,
Germantown, MD, USA) was used for 3D kinematic and kinetic variable computations. A
Cardan rotational sequence (X-y-z) was used for 3D angular kinematics computations and the
conventions of the angular kinematic and kinetic variables were defined using the right-hand
rule. Marker coordinate and GRF data were filtered using a fourth-order low-pass Butterworth
filter with cut-off frequency of 8 Hz for joint angle and moment computations. For GRF
variables, raw GRF data were filtered again using a fourth-order low-pass Butterworth filter with
cut-off frequency of 50 Hz. Joint moments were normalized to participant mass (Nm/kg) and
GRF variables were normalized to body weight (BW). Averages of selected variables across the
five trials for every condition for each participant were used in statistical analyses.
Using the Biodex Balance System SD software, overall, medial-lateral, and anterior-
posterior stability indices were calculated. Based on the center of balance point defined as [0,0]
of the X and Y coordinate of the system, X and Y coordinates values were recorded at each
sampling point (20 Hz). The X and Y coordinates effectively represent medial-lateral (X) and
anterior-posterior (Y) deviations from the center of balance. The Overall Stability Index (OSI)
was defined as:
177
𝑂𝑆𝐼 = √Σ(𝑋)2−Σ(𝑌)2
𝑁
where X is the medial-lateral coordinate, Y is the anterior-posterior coordinate, and N is the
number of data points sampled. The Medial-lateral Stability Index (MLSI) was calculated
through the following equation:
𝑀𝐿𝑆𝐼 = √Σ(𝑋)2
𝑁
where X is the medial-lateral coordinate and N is the number of data points sampled. The
Anterior-posterior Stability Index (APSI) was calculated as:
𝐴𝑃𝑆𝐼 = √Σ(𝑌)2
𝑁
where Y is the anterior-posterior coordinate, and N is the number of data points sampled.
A customized Matlab script (MathWorks, Natick, MA, USA) was created to calculate
strength related variables, including peak torque, angle of occurrence for peak torque, and
loading rate. Loading rate was calculated by dividing the peak torque by its time (from onset of
the movement).
Statistical Analyses
A correlation matrix was computed on all available variables to identify potential
variables for the regression model. If high correlations (r ≥ 0.7) existed, variables were selected
based on biomechanical and functional importance identified through review of literature
concerning TKR patients. Selected kinematic, kinetic, strength, and balance variables along with
functional test and survey data scores were input into a logistic regression analysis using SAS
(Version 9.4, Cary, NC, USA). Variables for inclusion in regression models were selected based
on three steps. First, a χ2 analysis was performed to determine the highest values of all variables.
Second, variables were chosen based on their biomechanical and clinical relevance and current
178
literature. Third, a correlation matrix was performed to reduce number of highly correlated
variables for redundancy, with the more relevant variable being selected for inclusion. Models
were evaluated using Akaike’s Information Criterion (AIC) values, comparing the nonnested
models of the same sample. The AIC value was calculated as:
𝐴𝐼𝐶 = −2𝐿𝑜𝑔𝐿 + 2𝑥
where x is the number of parameters in the model and calculated as:
𝑥 = 𝑟(𝑒 + 1)
where r is the total number of response levels and e is the number of explanatory effects. The
AIC uses a maximum log likelihood to add to the model parameters, calculated as:
−2 𝐿𝑜𝑔 𝐿 = −2 ∑𝑤𝑗
𝜎2𝑓𝑗log (�̂�𝑗)
𝑗
where w and f are weight and frequency values for the jth observation, and σ2 is the dispersion
parameter (253).
Goodness of fit is detailed with AIC and R2 values, taking model accuracy and
complexity into account. The model with the lowest AIC value was selected as the best model.
Significance levels were assessed using the Likelihood Ratio given the small sample size, with a
p level of 0.05. Using the Analysis of Maximum Likelihood Estimates, individual variable and
intercept slopes and standard errors were calculated. Odds ratios for the model variables were
calculated through the logistic regression, with the 95% Wald Confidence Limits. Models were
placed through two passes of the data, one without cross validation and then an additional test
with cross validation. This is performed by fitting the model to the entire data set. The model
was then used to predict probability for an observation by ignoring the selected observation and
fitting the model to the remainder of the observations to obtain a predicted probability for the
ignored observation. The data was not split into training and validation sets due to the small
179
sample size. Four main models were selected for further analysis. One for the more traditionally
assessed information with respect to patient satisfaction (survey, functional test, and strength
data) and three additional models: one for walking, stair ascent, and stair descent data (Table 13).
Results
The first model was inclusive of WOMAC total scores, stair ascent and chair rise times,
peak extension torque, and the Quadriceps/Hamstrings peak torque ratio, achieving the highest
R-square (0.87) and lowest AIC values (19.51, Table 13). Dissatisfied patients showed increased
WOMAC total scores, stair ascend times, and chair rise times and decreased peak knee extension
torques and quadriceps/hamstrings ratios. Individual variable and intercept slopes, standard
errors, odds ratios and confidence intervals are presented in Tables 14-17. None of the
individual variables showed significance for predictive ability in the model, despite the goodness
of fit.
The model on walking data achieved an R2 of 0.75 and an AIC of 21.34 (p=0.0015, Table
13). This model included 1st peak VGRF, 2nd peak VGRF, loading-response internal rotation
moments, and the FJS. Dissatisfied patients showed decreased 1st and 2nd peak VGRF values in
their replaced limb and peak loading-response internal rotation moments and increased FJS. A
second walking model was performed with the loading-response knee extension moment in place
of the internal rotation moment given that knee extension moments are more frequently
discussed with respect to TKR patients than internal rotation moments. The loading-response
internal rotation moment and knee extension moment were highly correlated (r=0.80) and the
internal rotation moment had a reduced χ2 value, resulting in the knee extension moment initially
being chosen for inclusion. However, upon further examination, when the internal rotation
moment was included in the model, the R-square increased by 0.06 and the AIC decreased by
180
1.40, resulting in a better model fit (both models were significant). Both walking models were
included for ease of literature comparison. No significant individual predictors were present
(p>0.05, Table 15).
The model on stair ascent data included 2nd peak VGRF, loading-response knee extension
moment, preferred ascent gait speed, and peak extension torque (180°/s), achieving an R2 of 0.72
and an AIC of 23.85 (p=0.0013). Dissatisfied patients showed reductions in all four variables,
with no significant individual predictors (p>0.05, Table 16). The model on stair descent data
included loading-response knee extension moments, preferred descent gait speed, peak knee
extension torque (180°/s), and FJS (R2=0.80, AIC=20.47, p=0.0003). Dissatisfied patients
showed reduced knee extension moments, preferred gait speed, and peak torque, with increased
FJS scores. No individual variables were significant predictors of patient dissatisfaction (p>0.05,
Table 17).
Discussion
The purpose of this research was to examine associations between different physical
characteristics and patient dissatisfaction after TKR operations in order to create four different
satisfaction regression models. In an examination of the different models, the model of best fit
contained survey, functional test, and strength data, achieving an R2 of 0.87 and an AIC value of
19.51 (Table 1). This model includes WOMAC total score, stair ascent and chair rise times,
peak knee extension torque, and the quadriceps to hamstring ratio. A prediction model for
patient dissatisfaction post TKR operation also included WOMAC total scores at one year post-
operation as a predictive variable (210). However, it should be noted that our research contains
more stringent inclusion and exclusion criteria, which may have altered the dissatisfaction
predictors. Dissatisfied patients reported higher WOMAC total scores in this data set, with
181
increased scores for pain, stiffness, and functional disability. While this model does not split the
WOMAC subsets, reduced functional abilities are evident in the model, as represented by the
increased stair ascent and chair rise times. Knee strength related measures have not been
previously included in predictive models for dissatisfied TKR patients prior to this research,
where reduced peak knee extension torques were evident for the dissatisfied patient population.
TKR patients exhibited a 331. Nm decrease in peak isokinetic knee extension torque compared
to healthy controls in the current study. This reduction in strength may have contributed to the
reduced functional ability of the dissatisfied patients. Lower knee extension power during
isokinetic testing (which is a product of torque and velocity) has been shown to be predictive of
slower stair ascent and descent tests (141). While power and torque are not directly related, peak
torque reductions were evident with reduced stair ascent times, both of which were included in
this model, suggesting the reduced functional ability being related to the reduced strength.
It was hypothesized that reduced knee extensor moments, reduced isokinetic quadriceps
strength, reduced gait speed, and increased pain would contribute to increased dissatisfaction in
TKR patients, which was partially confirmed by the walking, stair ascent, and stair descent
models. Reduced knee extension moments were present in the models for all three activities and
reduced quadriceps strength and preferred gait speed were evident in both stair ascent and
descent models. The two walking models were similar with respect to the variables entered into
them, however, the one with the higher R2 and lower AIC value included the peak loading-
response knee internal rotation moment. The model with slightly reduced fit statistics replaced
the internal rotation moment with the eventually chosen model with the peak loading-response
knee extension moment. The knee extension moment is a more important variable
biomechanically in gait and more frequently examined with respect to TKR patients (20, 24-26).
182
The dissatisfied patients exhibited reduced peak loading-response knee extension moments
compared to the satisfied patients. The dissatisfied patients showed reduced knee extension
strength in their replaced limb compared to their non-replaced limb and the satisfied patients. It
has been suggested that the reduced knee extension moment can be related to a quadriceps
avoidance gait pattern (25). By increasing the quadriceps strength and reducing the pain, the
dissatisfied patients may be able to increase knee loading. Additionally, both 1st and 2nd peak
VGRF levels were reduced in the dissatisfied group compared to the satisfied group. VGRF is
an indication of overall body loading but can also impact the joint moments, especially knee
extension moment, as is evidenced by simultaneous reductions in the knee extension moment
and VGRF levels for the dissatisfied patients. The final variable in this model was the FJS score,
with the dissatisfied patients showing increased scores on the FJS compared to satisfied patients.
The FJS was designed to examine the patient’s awareness of the replaced joint during different
activities, with the ability to forget about the replaced joint during activities of daily living as the
ultimate measure of satisfaction (158). This is a point not as evident in survey tools such as the
WOMAC. The increased scores for the dissatisfied patients indicate an increased awareness of
the joint and an inability for the joint to feel natural during movement. This may subsequently
reduce the functional abilities of the dissatisfied patients and increase the pain levels, both of
which are evident in the dissatisfied patients in this research as well as past research (219, 228).
The stair ascent model had two similar variables to the walking model, 2nd peak VGRF
and peak loading-response knee extension moment. Additionally, it was inclusive of preferred
ascent gait speed and peak knee extension torque at 180°/s. Reductions in all four variables were
present in the dissatisfied group compared to the satisfied group. Reductions in VGRF have
been seen in TKR patients compared to healthy controls during stair ascent at preferred gait
183
speed (235), but this was further compounded by the dissatisfied patients being lower than both
satisfied patients and healthy controls in 2nd peak VGRF and preferred gait speed. Knee
extension moments have mixed results in comparing TKR patients to healthy controls as reduced
knee extension moments have been reported by some studies (24, 39-41) while others have
reported no differences (22, 36, 42). As mentioned previously, reduced knee extension moments
are often due to weaker quadriceps muscles, leading to a quadriceps avoidance gait as a means of
avoiding knee joint loading (24, 129). This point is further accentuated by the reduced knee
extension strength and extension moments in the same predictive model. Satisfied TKR patients
had increased levels of all the included variables, suggesting that an increase in strength,
moments, and speed increases patient satisfaction.
The stair descent model was similar to the stair ascent model, with the inclusion of the
loading-response knee extension moment, preferred descent gait speed, and peak extension
torque (180°/s). The 2nd peak VGRF was replaced with the FJS. Dissatisfied patients showed
reduced knee extension moments, preferred gait speeds, and extension torque, while reporting
increased FJS scores. Research is mixed with respect to knee extension moments during stair
descent with one study showing reduced values in TKR patients compared to healthy controls
(22) while other studies have shown no differences (36, 40, 127). Reduced knee extension
moments have accompanied reduced gait velocity for TKR patients during stair descent (24, 39-
41). As with the stair ascent model, the reduction in quadriceps strength has also been linked to
knee extension moments and joint loading. Similar to the walking model, the presence of the
increased FJS score indicates an inability for dissatisfied patients to have reduced awareness of
the replaced joint. An increased awareness of the replaced joint may impact the patient’s ability
to perform daily activities using the joint.
184
In other investigations into satisfaction for TKR patients, pain has been consistently
reported as a factor for satisfaction prediction (18, 210, 254). None of the models presented in
this paper included specific VAS pain scores (aside from the WOMAC total score which has a
pain component to it). Pain was assessed after every test the TKR patients performed through
VAS, with the dissatisfied patients consistently reporting higher pain than the satisfied patients.
Based on initial analysis, pain would have been included in the models prior to WOMAC and
some additional variables as it had the higher Chi square values, therefore, based on the high
correlation of pain with the WOMAC, pain would have been input first. However when pain
levels were entered into any of the models, the model came back with an error message
indicating a complete separation of the data points. This essentially indicates that the presence of
pain in the dissatisfied patients is a universal truth, therefore the inclusion of pain in any model is
a perfect fit. While pain was not included in any model because of this, it is important to
highlight the effect pain has on the models and what its subsequent relationship to patient
dissatisfaction is.
This study has a few limitations worth mentioning. Most notably, the sample size is
much smaller than what traditional logistic regressions require. We only have nine dissatisfied
patients and fifteen satisfied patients. Despite this, to our knowledge, this is the first logistic
regression to analyze TKR patient dissatisfaction based on kinematic and kinetic variables.
Future research would benefit from a larger sample size of dissatisfied patients in order to add to
the power of the regression models, and confirm and strengthen these models. Additionally,
none of the included independent variables showed significance within the model, likely due to
the small sample size. Aside from this, model fit statistics were still significant and showed good
fit for the reported models. Finally, balance data was collected on this patient population,
185
however, due to an inability of many TKR patients to complete the unilateral balance tests, we
were unable to include the data in the model analyses. During static unilateral balance trials,
approximately 33% of the dissatisfied patients were unable to complete the trials and 13% of the
satisfied patients failed to do so. As for the dynamic unilateral balance trials, 56% of dissatisfied
patients could not complete the trials compared to 27% of the satisfied patients. This indicates a
reduced ability of the dissatisfied patients to complete balance trials, yet it means an inability of
the model to include the data due to lack of observations. In order to preserve the already small
sample size, we decided not to include the balance data. Future research would benefit from the
increased sample size for increased observations which may allow for inclusion of the balance
information.
Conclusion
As evidenced, patient satisfaction is a complicated construct. Through the use of
kinematic, kinetic, and strength data, a few variables were evident in creating models for
dissatisfaction prediction. Knee extension moments, VGRF, and knee isokinetic extension
strength were frequently evident for their contribution to patient dissatisfaction. Increasing the
patient’s ability to handle increases in these independent variables may help to improve their
satisfaction levels. Additionally, the presence of pain is the most important factor, despite not
being included in any of the models. Future research may benefit from examining improvements
in these independent variables and how that improves patient satisfaction rates.
186
Chapter VI Appendix: Tables and Figures
Table 13. Logistic regression models for TKR patient satisfaction with respect to survey data, strength, and 3D kinematics and kinetics for over
ground walking, stair ascent, and stair descent.
Model R2 AIC P value
WOMAC Total Score, Stair Ascend Time, Chair Rise Time, Peak Extension Torque (60°/s),
Quad/Hamstring Ratio (60°/s)
0.87 19.51 0.0002
Walking 1st Peak VGRF, Walking 2nd Peak VGRF, Walking Knee Internal Rotation Moment, FJS 0.75 21.34 0.0015
Walking 1st Peak VGRF, Walking 2nd Peak VGRF, Walking Knee Extension Moment, FJS 0.69 22.73 0.0026
Stair Ascend 2nd Peak VGRF, Ascend Loading Knee Extension Moment, Preferred Ascend gait speed,
Peak Extension torque (180°/s)
0.72 23.85 0.0013
Stair Descend Loading-response Knee Extension Moment, Preferred Descend gait speed, Peak
Extension Torque (180°/s), FJS
0.80 20.47 0.0003
Table 14. Slope (B), standard error (SE), Odds ratio (OR), and 95% confidence intervals (95% CI) for predictive variables for top model of
strength, functional test, and survey data.
Variable B SE P value OR 95% CI
WOMAC Total Score -0.06 0.04 0.1977 0.95 0.87 - 1.03
Stair Ascend Time -9.43 7.31 0.1970 <0.001 <0.001 to 133.75
Chair Rise Time 2.69 2.19 0.2186 14.72 0.20 to >999.99
Peak Extension Torque (60°/s) 0.01 0.05 0.8379 1.01 0.91 to 1.13
Quad/Hamstring Ratio (60°/s) 12.80 11.02 0.2454 >999.99 <0.001 to >999.99
Constant (Intercept) -1.28 8.13 0.8749 NA NA
187
Table 15. Slope (B), standard error (SE), Odds ratio (OR), and 95% confidence intervals (95% CI) for predictive variables for top model of
walking data.
Variable B SE P value OR 95% CI
1st Peak VGRF -30.65 41.56 0.4608 <0.001 <0.001 to >999.99
2nd Peak VGRF 43.56 48.94 0.3735 >999.99 <0.001 to >999.99
Loading-response Knee Extension Moment 10.92 10.48 0.2978 >999.99 <0.001 to >999.99
FJS 0.05 0.04 0.1743 1.05 0.98 to 1.13
Constant (Intercept) -18.89 17.03 0.2674 NA NA
Table 16. Slope (B), standard error (SE), Odds ratio (OR), and 95% confidence intervals (95% CI) for predictive variables for top model of stair
ascent data.
Variable B SE P value OR 95% CI
2nd Peak VGRF 17.74 10.93 0.1046 >999.99 0.025 to >999.99
Loading-response Knee Extension Moment 0.97 3.48 0.7801 2.64 0.003 to >999.99
Preferred Ascend Gait Speed 10.19 7.06 0.1491 >999.99 0.026 to >999.99
Peak Extension Torque (180°/s) 0.03 0.04 0.4162 1.03 0.96 to 1.10
Constant (Intercept) -28.60 13.21 0.0303 NA NA
Table 17. Slope (B), standard error (SE), Odds ratio (OR), and 95% confidence intervals (95% CI) for predictive variables for top model of stair
descent data.
Variable B SE P value OR 95% CI
Loading-response Knee Extension Moment -2.48 5.40 0.6461 0.08 <0.001 to >999.99
Preferred Descend Gait Speed 13.52 11.19 0.2272 >999.99 <0.001 to >999.99
Peak Extension Torque (180°/s) 0.09 0.06 0.1234 1.10 0.98 to 1.24
FJS 0.10 0.06 0.0756 1.10 0.99 to 1.23
Constant (Intercept) -16.20 9.10 0.0750 NA NA
188
CHAPTER VII
CONCLUSION
The purpose of this dissertation was to examine the strength, balance abilities, functional
abilities, and lower extremity biomechanics during level ground walking, stair ascent, and stair
descent of dissatisfied TKR patients. In study one, the analysis of overground walking
biomechanics showed that dissatisfied patients have altered level ground gait mechanics, with
the presence of increased knee joint pain and decreased preferred gait speed. Dissatisfied
patients exhibited reduced 1st and 2nd peak VGRF values, reduced knee flexion ROM, and knee
extension moments. The dissatisfied patients walk in such a manner that the knee joint
experiences reduced extension loads, likely as a means of alleviating the experienced pain and
joint loading. In study two, this was further evidenced during both stair ascent and descent
activities. The dissatisfied patients had reduced knee extension moments and abduction
moments, reduced preferred gait speed, and increased knee pain during ascent and descent
activities. Stair climbing is a more demanding activity compared to level ground walking, as is
evidenced by the amplified gait mechanics differences during these activities. The increased
changes may be the result of increasing levels of pain and thus a desire to unload the affected
joint.
In study three, the dissatisfied patients showed a reduced ability during functional exams
and reduced passive knee ROM, common assessments used during the rehabilitation process.
Reduced abilities were further evidenced by reduced strength levels and a more frequent inability
to complete single limb balance trials. Dissatisfied patients showed reduced knee extensor and
flexor strength, both of which are associated with deficits in gait mechanics. It is worth noting,
however, that when the ability to balance was present, there was no difference between satisfied
189
and dissatisfied TKR patients. In study four, a logistic regression analysis examined four
different models based on the data collected in this study, which reported good measures of fit in
prediction of patient dissatisfaction. These models often showed similarities in prediction
variables, such as the knee extension moment during overground walking, stair ascent, and stair
descent or one of the VGRF peaks (a measure of whole body loading).
The findings of this dissertation highlight that there are significant physical differences in
the dissatisfied TKR patients compared to satisfied TKR patients and healthy controls. With the
knee extension moment and knee extensor strength consistently being a source of significant
difference between satisfied and dissatisfied patients, it becomes evident that rehabilitation and
post-operative training should address this. Gait retraining may be an asset in order to teach
dissatisfied patients how to more appropriately handle the joint load. This may be further aided
through additional strength training as evidenced by the dissatisfied patients reduced strength.
Increased strength of the knee joint muscles may aid in dissatisfied patients ability to handle the
increased knee loading and potentially improve the satisfaction rates of TKR patients.
With every test performed, the dissatisfied patients reported increased levels of pain in
their replaced knee. When pain was entered into any of the regression models, a complete
separation of the data points existed, essentially implying that pain in the dissatisfied patients is a
truism. The presence of the knee pain is likely impairing knee mechanics, as there is a desire to
unload the joint when pain is present. As such, addressing pain is an additional pertinent issue
which may help to subsequently improve the mechanical factors of the knee joint. A goal of
TKR surgery is to alleviate joint pain experienced during osteoarthritis, however, in the case of
the dissatisfied patients, this goal was not met. From a physical standpoint, the pain should have
dissipated with the removal of the bone on bone contact present during end-stage osteoarthritis.
190
Given the continued presence of pain well over two years after surgery, these patients may
benefit from psychological training techniques which have been shown to help with pain
reduction, such as guided imagery, deep breathing, or relaxation techniques. These
characteristics all collectively illustrate the patient satisfaction is a multifaceted construct. Many
potential factors were not taken into consideration with this research (such as quality of life,
support, and other factors unrelated to their physical experiences). What this research does
highlight is that additional physical training during the rehabilitation process may be beneficial,
including mental skills training to help deal with the psychological aspects of increased joint
demands and pain. If the patients are unsure of the replaced knee’s ability to perform, it may
additionally lead to joint avoidance. Future research should examine a more holistic approach to
the rehabilitation process to examine the effects on patient satisfaction.
191
REFERENCES
192
1. Hootman, J.M. and C.G. Helmick, Projections of US prevalence of arthritis and
associated activity limitations. Arthritis and Rheumatism, 2006. 54(1): p. 226-229.
2. Statistics, N.C.f.H., Cost of hospital discharges with common hospital operating room
procedures in non-federal community hospitals, by age and selected principal procedure:
United States, selected years 2000-2011. 2013, Center for Disease Control and
Prevention: http://www.cdc.gov/nchs/hus/contents2012.htm.
3. Kurtz, S., et al., Projections of primary and revision hip and knee arthroplasty in the
United States from 2005 to 2030. Journal of Bone and Joint Surgery-American Volume,
2007. 89a(4): p. 780-785.
4. Ravi, B., et al., The changing demographics of total joint arthroplasty recipients in the
United States and Ontario from 2001 to 2007. Best Practice & Research in Clinical
Rheumatology, 2012. 26(5): p. 637-647.
5. Scuderi, G.R., et al., The New Knee Society Knee Scoring System. Clinical Orthopaedics
and Related Research, 2012. 470(1): p. 3-19.
6. Bruyere, O., et al., Health-related quality of life after total knee or hip replacement for
osteoarthritis: a 7-year prospective study. Archives of Orthopaedic and Trauma Surgery,
2012. 132(11): p. 1583-1587.
7. Hawker, G., et al., Health-related quality of life after knee replacement - Results of the
Knee Replacement Patient Outcomes Research Team Study. Journal of Bone and Joint
Surgery-American Volume, 1998. 80a(2): p. 163-173.
8. Ko, Y., et al., Health-Related Quality of Life after Total Knee Replacement or
Unicompartmental Knee Arthroplasty in an Urban Asian Population. Value in Health,
2011. 14(2): p. 322-328.
9. Kim, T.K., et al., Causes and Predictors of Patient's Dissatisfaction After Uncomplicated
Total Knee Arthroplasty. Journal of Arthroplasty, 2009. 24(2): p. 263-271.
10. Felts, E., et al., Function and quality of life following medial unicompartmental knee
arthroplasty in patients 60 years of age or younger. Orthop Traumatol Surg Res, 2010.
96(8): p. 861-7.
11. Weiss, J.M., et al., What functional activities are important to patients with knee
replacements? Clin Orthop Relat Res, 2002(404): p. 172-88.
12. Beswick, A.D., et al., What proportion of patients report long-term pain after total hip or
knee replacement for osteoarthritis? A systematic review of prospective studies in
unselected patients. BMJ Open, 2012. 2(1): p. e000435.
13. Noble, P.C., et al., Does total knee replacement restore normal knee function? Clin
Orthop Relat Res, 2005(431): p. 157-65.
14. Boonstra, M.C., M.C. De Waal Malefijt, and N. Verdonschot, How to quantify knee
function after total knee arthroplasty? Knee, 2008. 15(5): p. 390-5.
15. Ouellet, D. and H. Moffet, Locomotor deficits before and two months after knee
arthroplasty. Arthritis Rheum, 2002. 47(5): p. 484-93.
16. Ali, A., et al., Dissatisfied patients after total knee arthroplasty: a registry study
involving 114 patients with 8-13 years of followup. Acta Orthop, 2014. 85(3): p. 229-33.
17. Anderson, J.G., et al., Functional outcome and patient satisfaction in total knee patients
over the age of 75. Journal of Arthroplasty, 1996. 11(7): p. 831-840.
18. Baker, P.N., et al., The role of pain and function in determining patient satisfaction after
total knee replacement - Data from the National Joint Registry for England and Wales.
Journal of Bone and Joint Surgery-British Volume, 2007. 89b(7): p. 893-900.
193
19. Standifird, T.W., et al., Influence of Total Knee Arthroplasty on Gait Mechanics of the
Replaced and Non-Replaced Limb During Stair Negotiation. J Arthroplasty, 2016. 31(1):
p. 278-83.
20. Benedetti, M.G., et al., Muscle activation pattern and gait biomechanics after total knee
replacement. Clin Biomech (Bristol, Avon), 2003. 18(9): p. 871-6.
21. Metcalfe, A., et al., Abnormal loading of the major joints in knee osteoarthritis and the
response to knee replacement. Gait Posture, 2013. 37(1): p. 32-6.
22. Wilson, S.A., et al., Comprehensive gait analysis in posterior-stabilized knee
arthroplasty. J Arthroplasty, 1996. 11(4): p. 359-67.
23. McClelland, J.A., et al., Knee kinematics during walking at different speeds in people
who have undergone total knee replacement. Knee, 2011. 18(3): p. 151-5.
24. Mandeville, D., L.R. Osternig, and L.S. Chou, The effect of total knee replacement on
dynamic support of the body during walking and stair ascent. Clin Biomech (Bristol,
Avon), 2007. 22(7): p. 787-94.
25. Levinger, P., et al., Lower limb biomechanics in individuals with knee osteoarthritis
before and after total knee arthroplasty surgery. J Arthroplasty, 2013. 28(6): p. 994-9.
26. Smith, A.J., D.G. Lloyd, and D.J. Wood, A kinematic and kinetic analysis of walking
after total knee arthroplasty with and without patellar resurfacing. Clin Biomech
(Bristol, Avon), 2006. 21(4): p. 379-86.
27. Orishimo, K.F., et al., Does total knee arthroplasty change frontal plane knee
biomechanics during gait? Clin Orthop Relat Res, 2012. 470(4): p. 1171-6.
28. Mandeville, D., et al., The effect of total knee replacement on the knee varus angle and
moment during walking and stair ascent. Clin Biomech (Bristol, Avon), 2008. 23(8): p.
1053-8.
29. Nadeau, S., B.J. McFadyen, and F. Malouin, Frontal and sagittal plane analyses of the
stair climbing task in healthy adults aged over 40 years: what are the challenges
compared to level walking? Clin Biomech (Bristol, Avon), 2003. 18(10): p. 950-9.
30. Paquette, M.R., et al., Effects of increased step width on frontal plane knee biomechanics
in healthy older adults during stair descent. Knee, 2014. 21(4): p. 821-826.
31. Startzell, J.K., et al., Stair negotiation in older people: a review. J Am Geriatr Soc, 2000.
48(5): p. 567-80.
32. Williamson, J.D. and L.P. Fried, Characterization of older adults who attribute
functional decrements to "old age". J Am Geriatr Soc, 1996. 44(12): p. 1429-34.
33. Costigan, P.A., K.J. Deluzio, and U.P. Wyss, Knee and hip kinetics during normal stair
climbing. Gait Posture, 2002. 16(1): p. 31-7.
34. Dawson, J., et al., Questionnaire on the perceptions of patients about total knee
replacement. J Bone Joint Surg Br, 1998. 80(1): p. 63-9.
35. Insall, J.N., et al., Rationale of the Knee Society clinical rating system. Clin Orthop Relat
Res, 1989(248): p. 13-4.
36. Bolanos, A.A., et al., A comparison of isokinetic strength testing and gait analysis in
patients with posterior cruciate-retaining and substituting knee arthroplasties. J
Arthroplasty, 1998. 13(8): p. 906-15.
37. Kelman, G.J., et al., Gait laboratory analysis of a posterior cruciate-sparing total knee
arthroplasty in stair ascent and descent. Clin Orthop Relat Res, 1989(248): p. 21-5;
discussion 25-6.
194
38. Alnahdi, A.H., J.A. Zeni, and L. Snyder-Mackler, Gait after unilateral total knee
arthroplasty: frontal plane analysis. J Orthop Res, 2011. 29(5): p. 647-52.
39. Berti, L., et al., Clinical and biomechanical assessment of patella resurfacing in total
knee arthroplasty. Clin Biomech (Bristol, Avon), 2006. 21(6): p. 610-6.
40. Catani, F., et al., Mobile and fixed bearing total knee prosthesis functional comparison
during stair climbing. Clin Biomech (Bristol, Avon), 2003. 18(5): p. 410-8.
41. Fantozzi, S., et al., Fluoroscopic and gait analysis of the functional performance in stair
ascent of two total knee replacement designs. Gait Posture, 2003. 17(3): p. 225-34.
42. Saari, T., et al., Total knee replacement influences both knee and hip joint kinematics
during stair climbing. Int Orthop, 2004. 28(2): p. 82-6.
43. Mizner, R.L., et al., Preoperative quadriceps strength predicts functional ability one year
after total knee arthroplasty. J Rheumatol, 2005. 32(8): p. 1533-9.
44. Judd, D.L., D.G. Eckhoff, and J.E. Stevens-Lapsley, Muscle strength loss in the lower
limb after total knee arthroplasty. Am J Phys Med Rehabil, 2012. 91(3): p. 220-6; quiz
227-30.
45. Schroer, W.C., et al., Isokinetic strength testing of minimally invasive total knee
arthroplasty recovery. J Arthroplasty, 2010. 25(2): p. 274-9.
46. Yoshida, Y., et al., Examining outcomes from total knee arthroplasty and the relationship
between quadriceps strength and knee function over time. Clin Biomech (Bristol, Avon),
2008. 23(3): p. 320-8.
47. Swinkels, A., J.H. Newman, and T.J. Allain, A prospective observational study of falling
before and after knee replacement surgery. Age Ageing, 2009. 38(2): p. 175-81.
48. Gauchard, G.C., et al., On the role of knee joint in balance control and postural
strategies: effects of total knee replacement in elderly subjects with knee osteoarthritis.
Gait Posture, 2010. 32(2): p. 155-60.
49. Liao, C.D., et al., Functional outcomes of outpatient balance training following total knee
replacement in patients with knee osteoarthritis: a randomized controlled trial. Clin
Rehabil, 2015. 29(9): p. 855-67.
50. Pua, Y.H., et al., Associations of knee extensor strength and standing balance with
physical function in knee osteoarthritis. Arthritis Care Res (Hoboken), 2011. 63(12): p.
1706-14.
51. Cho, S.D. and C.H. Hwang, Improved single-limb balance after total knee arthroplasty.
Knee Surg Sports Traumatol Arthrosc, 2013. 21(12): p. 2744-50.
52. Stensdotter, A.K., J. Bjerke, and M. Djupsjobacka, Postural sway in single-limb and
bilateral quiet standing after unilateral total knee arthroplasty. Gait Posture, 2015. 41(3):
p. 769-73.
53. Hatfield, G.L., et al., The effect of total knee arthroplasty on knee joint kinematics and
kinetics during gait. J Arthroplasty, 2011. 26(2): p. 309-18.
54. Mandeville, D., L.R. Osternig, and L.S. Chou, The effect of total knee replacement
surgery on gait stability. Gait Posture, 2008. 27(1): p. 103-9.
55. White, R., et al., The variability of force platform data in normal and cerebral palsy gait.
Clin Biomech (Bristol, Avon), 1999. 14(3): p. 185-92.
56. Sigal, L., A.O. Balan, and M.J. Black, HUMANEVA: Synchronized Video and Motion
Capture Dataset and Baseline Algorithm for Evaluation of Articulated Human Motion.
International Journal of Computer Vision, 2010. 87: p. 4-27.
195
57. Winter, D.A., Biomechanics and motor control of human movement. 2009: John Wiley &
Sons.
58. Plankers, R. and P. Fua, Tracking and modeling people in video sequences. Computer
Vision and Image Understanding, 2001. 81: p. 285-302.
59. Liu, J. and T.E. Lockhart, Comparison of 3D joint moments using local and global
inverse dynamics approaches among three different age groups. Gait Posture, 2006.
23(4): p. 480-5.
60. Akbarshahi, M., et al., Non-invasive assessment of soft-tissue artifact and its effect on
knee joint kinematics during functional activity. J Biomech, 2010. 43(7): p. 1292-301.
61. Kuo, M.Y., et al., Influence of soft tissue artifacts on the calculated kinematics and
kinetics of total knee replacements during sit-to-stand. Gait Posture, 2011. 33(3): p. 379-
84.
62. Alkjaer, T., E.B. Simonsen, and P. Dyhre-Poulsen, Comparison of inverse dynamics
calculated by two- and three-dimensional models during walking. Gait Posture, 2001.
13(2): p. 73-7.
63. Davis, R.B., Clinical gait analysis. IEEE Eng Med Biol Mag, 1988. 7(3): p. 35-40.
64. Collins, T.D., et al., A six degrees-of-freedom marker set for gait analysis: repeatability
and comparison with a modified Helen Hayes set. Gait Posture, 2009. 30(2): p. 173-80.
65. Schwartz, M.H., J.P. Trost, and R.A. Wervey, Measurement and management of errors in
quantitative gait data. Gait Posture, 2004. 20(2): p. 196-203.
66. Moeslund, T.B. and E. Granum, A survey of computer vision-based human motion
capture. Computer Vision and Image Understanding, 2001. 81: p. 231-268.
67. Weinhandl, J.T., et al., Validation of a single camera three-dimensional motion tracking
system. J Biomech, 2010. 43(7): p. 1437-40.
68. Miro, F., et al., 2D versus 3D in the kinematic analysis of the horse at the trot. Vet Res
Commun, 2009. 33(6): p. 507-13.
69. Kadaba, M.P., H.K. Ramakrishnan, and M.E. Wootten, Measurement of lower extremity
kinematics during level walking. J Orthop Res, 1990. 8(3): p. 383-92.
70. Cappozzo, A., Gait Analysis Methodology. Human Movement Science, 1984. 3(1-2): p.
27-50.
71. Alexander, E.J. and T.P. Andriacchi, Correcting for deformation in skin-based marker
systems. J Biomech, 2001. 34(3): p. 355-61.
72. Andersen, M.S., et al., Do kinematic models reduce the effects of soft tissue artefacts in
skin marker-based motion analysis? An in vivo study of knee kinematics. J Biomech,
2010. 43(2): p. 268-73.
73. Benoit, D.L., et al., Effect of skin movement artifact on knee kinematics during gait and
cutting motions measured in vivo. Gait Posture, 2006. 24(2): p. 152-64.
74. Holden, J.P., et al., Surface movement errors in shank kinematics and knee kinetics
during gait. Gait Posture, 1997. 5: p. 217-227.
75. Manal, K., et al., Knee moment profiles during walking: errors due to soft tissue
movement of the shank and the influence of the reference coordinate system. Gait
Posture, 2002. 15(1): p. 10-7.
76. Ramsey, D.K., et al., Methodological concerns using intra-cortical pins to measure
tibiofemoral kinematics. Knee Surg Sports Traumatol Arthrosc, 2003. 11(5): p. 344-9.
77. Reinschmidt, C., et al., Tibiofemoral and tibiocalcaneal motion during walking: external
vs. skeletal markers. Gait Posture, 1997. 6: p. 98-109.
196
78. Ryu, T., H.S. Choi, and M.K. Chung, Soft tissue artifact compensation using
displacement dependency between anatomical landmarks and skin markers--a
preliminary study. International Journal of Industrial Ergonomics, 2009. 39: p. 152-158.
79. Della Croce, U., et al., Human movement analysis using stereophotogrammetry: Part 4:
assessment of anatomical landmark misplacement and its effects on joint kinematics. Gait
Posture, 2005. 21(2): p. 226-237.
80. Manal, K., et al., Comparison of surface mounted markers and attachment methods in
estimating tibial rotations during walking: an in vivo study. Gait Posture, 2000. 11(1): p.
38-45.
81. Ferrari, A., et al., Quantitative comparison of five current protocols in gait analysis. Gait
Posture, 2008. 28(2): p. 207-16.
82. Gao, B. and N.N. Zheng, Investigation of soft tissue movement during level walking:
translations and rotations of skin markers. J Biomech, 2008. 41(15): p. 3189-95.
83. Miranda, D.L., et al., Kinematic differences between optical motion capture and biplanar
videoradiography during a jump-cut maneuver. J Biomech, 2013. 46(3): p. 567-73.
84. Stagni, R., et al., Quantification of soft tissue artefact in motion analysis by combining
3D fluoroscopy and stereophotogrammetry: a study on two subjects. Clin Biomech
(Bristol, Avon), 2005. 20(3): p. 320-9.
85. Tsai, T.Y., et al., Effects of soft tissue artifacts on the calculated kinematics and kinetics
of the knee during stair-ascent. J Biomech, 2011. 44(6): p. 1182-8.
86. Cappozzo, A., et al., Position and orientation in space of bones during movement:
experimental artefacts. Clin Biomech (Bristol, Avon), 1996. 11(2): p. 90-100.
87. Cappello, A., et al., Multiple anatomical landmark calibration for optimal bone pose
estimation. Human Movement Science, 1997. 16: p. 259-274.
88. Andriacchi, T.P., et al., A point cluster method for in vivo motion analysis: applied to a
study of knee kinematics. J Biomech Eng, 1998. 120(6): p. 743-9.
89. Perry, J. and J.R. Davids, Gait analysis: normal and pathological function. Journal of
Pediatric Orthopaedics, 1992. 12(6): p. 815.
90. Gorton, G.E., 3rd, D.A. Hebert, and M.E. Gannotti, Assessment of the kinematic
variability among 12 motion analysis laboratories. Gait Posture, 2009. 29(3): p. 398-402.
91. McGinley, J.L., et al., The reliability of three-dimensional kinematic gait measurements:
a systematic review. Gait Posture, 2009. 29(3): p. 360-9.
92. Corazza, S., et al., A markerless motion capture system to study musculoskeletal
biomechanics: visual hull and simulated annealing approach. Ann Biomed Eng, 2006.
34(6): p. 1019-29.
93. Chao, E.Y., et al., Normative data of knee joint motion and ground reaction forces in
adult level walking. J Biomech, 1983. 16(3): p. 219-33.
94. Soderkvist, I. and P.A. Wedin, Determining the movements of the skeleton using well-
configured markers. J Biomech, 1993. 26(12): p. 1473-7.
95. Grood, E.S. and W.J. Suntay, A joint coordinate system for the clinical description of
three-dimensional motions: application to the knee. J Biomech Eng, 1983. 105(2): p.
136-44.
96. Wu, G., et al., ISB recommendation on definitions of joint coordinate system of various
joints for the reporting of human joint motion--part I: ankle, hip, and spine. International
Society of Biomechanics. J Biomech, 2002. 35(4): p. 543-8.
197
97. Schache, A.G., R. Baker, and C.L. Vaughan, Differences in lower limb transverse plane
joint moments during gait when expressed in two alternative reference frames. J
Biomech, 2007. 40(1): p. 9-19.
98. Kristianslund, E., et al., Expressing the joint moments of drop jumps and sidestep cutting
in different reference frames--does it matter? J Biomech, 2014. 47(1): p. 193-9.
99. Woltring, H.J., Representation and Calculation of 3-D Joint Movement. Human
Movement Science, 1991. 10(5): p. 603-616.
100. Cole, G.K., et al., Application of the Joint Coordinate System to 3-Dimensional Joint
Attitude and Movement Representation - a Standardization Proposal. Journal of
Biomechanical Engineering-Transactions of the Asme, 1993. 115(4): p. 344-349.
101. Lees, A., G. Barton, and M. Robinson, The influence of Cardan rotation sequence on
angular orientation data for the lower limb in the soccer kick. Journal of Sports Sciences,
2010. 28(4): p. 445-450.
102. Senk, M. and L. Cheze, Rotation sequence as an important factor in shoulder kinematics.
Clinical Biomechanics, 2006. 21: p. S3-S8.
103. Cappozzo, A., et al., Human movement analysis using stereophotogrammetry - Part 1:
theoretical background. Gait & Posture, 2005. 21(2): p. 186-196.
104. Choisne, J., et al., Influence of kinematic analysis methods on detecting ankle and
subtalar joint instability. Journal of Biomechanics, 2012. 45(1): p. 46-52.
105. Schache, A.G. and R. Baker, On the expression of joint moments during gait. Gait
Posture, 2007. 25(3): p. 440-52.
106. Schache, A.G., R. Baker, and L.W. Lamoreux, Defining the knee joint flexion-extension
axis for purposes of quantitative gait analysis: an evaluation of methods. Gait Posture,
2006. 24(1): p. 100-9.
107. Barbour, K., et al., Prevalence of doctor-diagnosed arthritis and arthritis-attributable
activity limitation--United States, 2010-2012. Morbidity and Mortality Weekly Report,
2013. 62(44): p. 869-873.
108. Mehrotra, C., et al., Trends in total knee replacement surgeries and implications for
public health, 1990-2000. Public Health Reports, 2005. 120(3): p. 278-282.
109. Paquette, M.R., et al., Does increasing step width alter knee biomechanics in medial
compartment knee osteoarthritis patients during stair descent? Knee, 2014. 21(3): p.
676-82.
110. Tekin, B., B. Unver, and V. Karatosun, Expectations in patients with total knee
arthroplasty. Acta Orthop Traumatol Turc, 2012. 46(3): p. 174-80.
111. Noble, P.C., et al., The John Insall Award: Patient expectations affect satisfaction with
total knee arthroplasty. Clin Orthop Relat Res, 2006. 452: p. 35-43.
112. Parvizi, J., et al., High level of residual symptoms in young patients after total knee
arthroplasty. Clin Orthop Relat Res, 2014. 472(1): p. 133-7.
113. Kramers-de Quervain, I.A., et al., Prospective study of gait function before and 2 years
after total knee arthroplasty. Knee, 2012. 19(5): p. 622-7.
114. Milner, C.E. and M.E. O'Bryan, Bilateral frontal plane mechanics after unilateral total
knee arthroplasty. Arch Phys Med Rehabil, 2008. 89(10): p. 1965-9.
115. Ainsworth, B.E., et al., 2011 Compendium of Physical Activities: a second update of
codes and MET values. Med Sci Sports Exerc, 2011. 43(8): p. 1575-81.
116. Turcot, K., et al., How gait and clinical outcomes contribute to patients' satisfaction three
months following a total knee arthroplasty. J Arthroplasty, 2013. 28(8): p. 1297-300.
198
117. Protopapadaki, A., et al., Hip, knee, ankle kinematics and kinetics during stair ascent and
descent in healthy young individuals. Clin Biomech (Bristol, Avon), 2007. 22(2): p. 203-
10.
118. McMahon, M. and J.A. Block, The risk of contralateral total knee arthroplasty after knee
replacement for osteoarthritis. J Rheumatol, 2003. 30(8): p. 1822-4.
119. Hamel, K.A., et al., A comparison of foot/ground interaction during stair negotiation and
level walking in young and older women. Ergonomics, 2005. 48(8): p. 1047-56.
120. Standifird, T.W., Lower Extremity Biomechanics during Stair Ascent in Healthy and
Total Knee Replacement Older Adults, in Kinesiology, Recreation, and Sports Studies.
2015, University of Tennessee Knoxville: Trace: Tennessee Research and Creative
Exchange. p. 231.
121. Morrison, J.B., Function of the knee joint in various activities. Biomed Eng, 1969. 4(12):
p. 573-80.
122. Rasnick, R., et al., Knee Joint Loads and Surrounding Muscle Forces during Stair Ascent
in Patients with Total Knee Replacement. PLoS One, 2016. 11(6): p. e0156282.
123. Luepongsak, N., et al., The contribution of type of daily activity to loading across the hip
and knee joints in the elderly. Osteoarthritis Cartilage, 2002. 10(5): p. 353-9.
124. Milner, C.E., Is gait normal after total knee arthroplasty? Systematic review of the
literature. J Orthop Sci, 2009. 14(1): p. 114-20.
125. Jevsevar, D.S., et al., Knee kinematics and kinetics during locomotor activities of daily
living in subjects with knee arthroplasty and in healthy control subjects. Phys Ther, 1993.
73(4): p. 229-39; discussion 240-2.
126. Pollo, F.E., et al., Walking, chair rising, and stair climbing after total knee arthroplasty:
patellar resurfacing versus nonresurfacing. Am J Knee Surg, 2000. 13(2): p. 103-8;
discussion 108-9.
127. Joglekar, S., et al., Gait analysis comparison of cruciate retaining and substituting TKA
following PCL sacrifice. Knee, 2012. 19(4): p. 279-85.
128. McClelland, J.A., K.E. Webster, and J.A. Feller, Variability of walking and other daily
activities in patients with total knee replacement. Gait Posture, 2009. 30(3): p. 288-95.
129. Andriacchi, T.P., J.O. Galante, and R.W. Fermier, The influence of total knee-
replacement design on walking and stair-climbing. J Bone Joint Surg Am, 1982. 64(9): p.
1328-35.
130. Milner, C.E., Interlimb asymmetry during walking following unilateral total knee
arthroplasty. Gait Posture, 2008. 28(1): p. 69-73.
131. Andriacchi, T.P., et al., A study of lower-limb mechanics during stair-climbing. J Bone
Joint Surg Am, 1980. 62(5): p. 749-57.
132. Standifird, T.W., H.E. Cates, and S. Zhang, Stair ambulation biomechanics following
total knee arthroplasty: a systematic review. J Arthroplasty, 2014. 29(9): p. 1857-62.
133. Zhao, D., et al., Correlation between the knee adduction torque and medial contact force
for a variety of gait patterns. J Orthop Res, 2007. 25(6): p. 789-97.
134. Baliunas, A.J., et al., Increased knee joint loads during walking are present in subjects
with knee osteoarthritis. Osteoarthritis Cartilage, 2002. 10(7): p. 573-9.
135. Hurwitz, D.E., et al., The knee adduction moment during gait in subjects with knee
osteoarthritis is more closely correlated with static alignment than radiographic disease
severity, toe out angle and pain. J Orthop Res, 2002. 20(1): p. 101-7.
199
136. Miyazaki, T., et al., Dynamic load at baseline can predict radiographic disease
progression in medial compartment knee osteoarthritis. Ann Rheum Dis, 2002. 61(7): p.
617-22.
137. Bennett, H.J., Static Frontal Plane Tibiofemoral Alignments and Their Effects on Knee
Biomechanics During Level Walking and Stair Ascent Tasks, in Kinesiology, Recreation,
and Sports Studies. 2016, University of Tennessee Knoxville: Trace: Tennessee Research
and Creative Exchange. p. 194.
138. Debbi, E.M., et al., Frontal plane biomechanics of the operated and non-operated knees
before and after unilateral total knee arthroplasty. Clin Biomech (Bristol, Avon), 2015.
30(9): p. 889-94.
139. Vahtrik, D., et al., Postural stability in relation to anthropometric and functional
characteristics in women with knee osteoarthritis following total knee arthroplasty. Arch
Orthop Trauma Surg, 2014. 134(5): p. 685-92.
140. Huang, C.H., et al., Muscle strength after successful total knee replacement: a 6- to 13-
year followup. Clin Orthop Relat Res, 1996(328): p. 147-54.
141. Valtonen, A., et al., Muscle deficits persist after unilateral knee replacement and have
implications for rehabilitation. Phys Ther, 2009. 89(10): p. 1072-9.
142. Rossi, M.D., et al., Comparison of knee extensor strength between limbs in individuals
with bilateral total knee replacement. Arch Phys Med Rehabil, 2002. 83(4): p. 523-6.
143. Vandekerckhove, P.J., et al., Does cruciate retention primary total knee arthroplasty
affect proprioception, strength and clinical outcome? Knee Surg Sports Traumatol
Arthrosc, 2015. 23(6): p. 1644-52.
144. Franklin, P.D., W. Li, and D.C. Ayers, The Chitranjan Ranawat Award: functional
outcome after total knee replacement varies with patient attributes. Clin Orthop Relat
Res, 2008. 466(11): p. 2597-604.
145. Topp, R., et al., The effect of prehabilitation exercise on strength and functioning after
total knee arthroplasty. PM R, 2009. 1(8): p. 729-35.
146. Rodgers, J.A., et al., Preoperative physical therapy in primary total knee arthroplasty. J
Arthroplasty, 1998. 13(4): p. 414-21.
147. Rahmann, A.E., S.G. Brauer, and J.C. Nitz, A specific inpatient aquatic physiotherapy
program improves strength after total hip or knee replacement surgery: a randomized
controlled trial. Arch Phys Med Rehabil, 2009. 90(5): p. 745-55.
148. Lorentzen, J.S., et al., Early changes in muscle strength after total knee arthroplasty. A 6-
month follow-up of 30 knees. Acta Orthop Scand, 1999. 70(2): p. 176-9.
149. Gstoettner, M., et al., Preoperative proprioceptive training in patients with total knee
arthroplasty. Knee, 2011. 18(4): p. 265-70.
150. Doyle, T.L., R.U. Newton, and A.F. Burnett, Reliability of traditional and fractal
dimension measures of quiet stance center of pressure in young, healthy people. Arch
Phys Med Rehabil, 2005. 86(10): p. 2034-40.
151. Alviar, M.J., et al., Do Patient-Reported Outcome Measures in Hip and Knee
Arthroplasty Rehabilitation Have Robust Measurement Attributes? A Systematic Review.
Journal of Rehabilitation Medicine, 2011. 43(7): p. 572-583.
152. Bellamy, N. and W.W. Buchanan, A preliminary evaluation of the dimensionality and
clinical importance of pain and disability in osteoarthritis of the hip and knee. Clin
Rheumatol, 1986. 5(2): p. 231-41.
200
153. Dawson, J., et al., Comparison of measures to assess outcomes in total hip replacement
surgery. Qual Health Care, 1996. 5(2): p. 81-8.
154. Ware, J.E., Jr. and C.D. Sherbourne, The MOS 36-item short-form health survey (SF-36).
I. Conceptual framework and item selection. Med Care, 1992. 30(6): p. 473-83.
155. Terwee, C.B., et al., Quality criteria were proposed for measurement properties of health
status questionnaires. J Clin Epidemiol, 2007. 60(1): p. 34-42.
156. Streiner, D.L. and G.R. Norman, Health measurement scales: a practical guide to their
development and use. 1995, Oxford: Oxford University Press.
157. Dunbar, M.J., et al., Appropriate questionnaires for knee arthroplasty. Results of a survey
of 3600 patients from The Swedish Knee Arthroplasty Registry. J Bone Joint Surg Br,
2001. 83(3): p. 339-44.
158. Behrend, H., et al., The "forgotten joint" as the ultimate goal in joint arthroplasty:
validation of a new patient-reported outcome measure. J Arthroplasty, 2012. 27(3): p.
430-436 e1.
159. Nunnally, J.C., Psychometric theory. 1978, New York: McGraw-Hill Inc.
160. de Vet, H.C., et al., Are factor analytical techniques used appropriately in the validation
of health status questionnaires? A systematic review on the quality of factor analysis of
the SF-36. Qual Life Res, 2005. 14(5): p. 1203-18; dicussion 1219-21, 1223-4.
161. Charoencholvanich, K. and B. Pongcharoen, Oxford knee score and SF-36: translation &
reliability for use with total knee arthroscopy patients in Thailand. J Med Assoc Thai,
2005. 88(9): p. 1194-202.
162. Brazier, J.E., et al., Generic and condition-specific outcome measures for people with
osteoarthritis of the knee. Rheumatology (Oxford), 1999. 38(9): p. 870-7.
163. Gabr, A., R. Tansey, and F.S. Haddad, Outcome Measures in Total Knee Arthroplasty.
2015, Switzerland: Springer International Publishing. 79-87.
164. de Vet, H.C., et al., Reproducibility and responsiveness of evaluative outcome measures.
Theoretical considerations illustrated by an empirical example. Int J Technol Assess
Health Care, 2001. 17(4): p. 479-87.
165. Norman, G.R., J.A. Sloan, and K.W. Wyrwich, Interpretation of changes in health-
related quality of life: the remarkable universality of half a standard deviation. Med
Care, 2003. 41(5): p. 582-92.
166. Angst, F., et al., Responsiveness of the WOMAC osteoarthritis index as compared with
the SF-36 in patients with osteoarthritis of the legs undergoing a comprehensive
rehabilitation intervention. Ann Rheum Dis, 2001. 60(9): p. 834-40.
167. de Vet, H.C., et al., When to use agreement versus reliability measures. J Clin Epidemiol,
2006. 59(10): p. 1033-9.
168. Bland, J.M. and D.G. Altman, Statistical methods for assessing agreement between two
methods of clinical measurement. Lancet, 1986. 1(8476): p. 307-10.
169. Terwee, C.B., et al., On assessing responsiveness of health-related quality of life
instruments: guidelines for instrument evaluation. Qual Life Res, 2003. 12(4): p. 349-62.
170. Osoba, D., et al., Interpreting the significance of changes in health-related quality-of-life
scores. J Clin Oncol, 1998. 16(1): p. 139-44.
171. Jaeschke, R., J. Singer, and G.H. Guyatt, Measurement of health status. Ascertaining the
minimal clinically important difference. Control Clin Trials, 1989. 10(4): p. 407-15.
172. Blanchard, C., et al., Is the Health Utilities Index responsive in total hip arthroplasty
patients? J Clin Epidemiol, 2003. 56(11): p. 1046-54.
201
173. Crosby, R.D., R.L. Kolotkin, and G.R. Williams, Defining clinically meaningful change
in health-related quality of life. J Clin Epidemiol, 2003. 56(5): p. 395-407.
174. Dunbar, M.J., et al., Translation and validation of the Oxford-12 item knee score for use
in Sweden. Acta Orthop Scand, 2000. 71(3): p. 268-74.
175. Wright, R.W., Knee injury outcomes measures. J Am Acad Orthop Surg, 2009. 17(1): p.
31-9.
176. McConnell, S., P. Kolopack, and A.M. Davis, The Western Ontario and McMaster
Universities Osteoarthritis Index (WOMAC): a review of its utility and measurement
properties. Arthritis Rheum, 2001. 45(5): p. 453-61.
177. Bellamy, N., et al., Validation study of WOMAC: a health status instrument for
measuring clinically important patient-relevant outcomes following total hip or knee
arthroplasty in osteoarthritis. J Orthop Rheumatol, 1988. 1: p. 95-108.
178. Bellamy, N., et al., Validation study of WOMAC: a health status instrument for
measuring clinically important patient relevant outcomes to antirheumatic drug therapy
in patients with osteoarthritis of the hip or knee. J Rheumatol, 1988. 15(12): p. 1833-40.
179. Ostendorf, M., et al., Patient-reported outcome in total hip replacement. A comparison of
five instruments of health status. J Bone Joint Surg Br, 2004. 86(6): p. 801-8.
180. Soderman, P. and H. Malchau, Validity and reliability of Swedish WOMAC osteoarthritis
index: a self-administered disease-specific questionnaire (WOMAC) versus generic
instruments (SF-36 and NHP). Acta Orthop Scand, 2000. 71(1): p. 39-46.
181. Xie, F., et al., Validation of Chinese Western Ontario and McMaster Universities
Osteoarthritis Index (WOMAC) in patients scheduled for total knee replacement. Qual
Life Res, 2008. 17(4): p. 595-601.
182. Kennedy, D., et al., Exploring the factorial validity and clinical interpretability of the
Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). Physiother
Can, 2003. 55: p. 160-168.
183. Roorda, L.D., et al., Satisfactory cross cultural equivalence of the Dutch WOMAC in
patients with hip osteoarthritis waiting for arthroplasty. Ann Rheum Dis, 2004. 63(1): p.
36-42.
184. Stratford, P.W., et al., Measurement properties of the WOMAC LK 3.1 pain scale.
Osteoarthritis Cartilage, 2007. 15(3): p. 266-72.
185. Chesworth, B.M., et al., Willingness to go through surgery again validated the WOMAC
clinically important difference from THR/TKR surgery. J Clin Epidemiol, 2008. 61(9): p.
907-18.
186. Escobar, A., et al., Responsiveness and clinically important differences for the WOMAC
and SF-36 after total knee replacement. Osteoarthritis Cartilage, 2007. 15(3): p. 273-80.
187. Quintana, J.M., et al., Responsiveness and clinically important differences for the
WOMAC and SF-36 after hip joint replacement. Osteoarthritis Cartilage, 2005. 13(12): p.
1076-83.
188. Bachmeier, C.J., et al., A comparison of outcomes in osteoarthritis patients undergoing
total hip and knee replacement surgery. Osteoarthritis Cartilage, 2001. 9(2): p. 137-46.
189. Haverkamp, D., et al., Translation and validation of the Dutch version of the Oxford 12-
item knee questionnaire for knee arthroplasty. Acta Orthop, 2005. 76(3): p. 347-52.
190. Matsumoto, M., et al., Validation study of the Forgotten Joint Score-12 as a universal
patient-reported outcome measure. Eur J Orthop Surg Traumatol, 2015. 25(7): p. 1141-5.
202
191. Thienpont, E., et al., Joint awareness in different types of knee arthroplasty evaluated
with the Forgotten Joint score. J Arthroplasty, 2014. 29(1): p. 48-51.
192. Hamilton, D.F., et al., What determines patient satisfaction with surgery? A prospective
cohort study of 4709 patients following total joint replacement. Bmj Open, 2013. 3(4).
193. Impellizzeri, F.M., et al., Comparison of the Reliability, Responsiveness, and Construct
Validity of 4 Different Questionnaires for Evaluating Outcomes after Total Knee
Arthroplasty. Journal of Arthroplasty, 2011. 26(6): p. 861-869.
194. Robertsson, O. and M.J. Dunbar, Patient satisfaction compared with general health and
disease-specific questionnaires in knee arthroplasty patients. J Arthroplasty, 2001. 16(4):
p. 476-82.
195. Wylde, V., et al., Patient-reported outcomes after total hip and knee arthroplasty:
comparison of midterm results. J Arthroplasty, 2009. 24(2): p. 210-6.
196. Xie, F., et al., Cross-cultural adaptation and validation of Singapore English and
Chinese Versions of the Oxford Knee Score (OKS) in knee osteoarthritis patients
undergoing total knee replacement. Osteoarthritis Cartilage, 2007. 15(9): p. 1019-24.
197. Roos, E.M., et al., Knee Injury and Osteoarthritis Outcome Score (KOOS)--development
of a self-administered outcome measure. J Orthop Sports Phys Ther, 1998. 28(2): p. 88-
96.
198. Roos, E.M. and S. Toksvig-Larsen, Knee injury and Osteoarthritis Outcome Score
(KOOS) - validation and comparison to the WOMAC in total knee replacement. Health
Qual Life Outcomes, 2003. 1: p. 17.
199. Davis, A.M., et al., Comparative, validity and responsiveness of the HOOS-PS and
KOOS-PS to the WOMAC physical function subscale in total joint replacement for
osteoarthritis. Osteoarthritis Cartilage, 2009. 17(7): p. 843-7.
200. de Groot, I.B., et al., The Dutch version of the Knee Injury and Osteoarthritis Outcome
Score: a validation study. Health Qual Life Outcomes, 2008. 6: p. 16.
201. Bombardier, C., et al., Comparison of a generic and a disease-specific measure of pain
and physical function after knee replacement surgery. Med Care, 1995. 33(4 Suppl): p.
AS131-44.
202. Becker, R., et al., Expectation, satisfaction and clinical outcome of patients after total
knee arthroplasty. Knee Surg Sports Traumatol Arthrosc, 2011. 19(9): p. 1433-41.
203. Conditt, M.A., et al., The PCL significantly affects the functional outcome of total knee
arthroplasty. J Arthroplasty, 2004. 19(7 Suppl 2): p. 107-12.
204. Ware, J.E., Jr., et al., Comparison of methods for the scoring and statistical analysis of
SF-36 health profile and summary measures: summary of results from the Medical
Outcomes Study. Med Care, 1995. 33(4 Suppl): p. AS264-79.
205. Ware, J., Jr., M. Kosinski, and S.D. Keller, A 12-Item Short-Form Health Survey:
construction of scales and preliminary tests of reliability and validity. Med Care, 1996.
34(3): p. 220-33.
206. Stucki, G., et al., The Short Form-36 is preferable to the SIP as a generic health status
measure in patients undergoing elective total hip arthroplasty. Arthritis Care Res, 1995.
8(3): p. 174-81.
207. Busija, L., et al., Magnitude and meaningfulness of change in SF-36 scores in four types
of orthopedic surgery. Health Qual Life Outcomes, 2008. 6: p. 55.
208. Hossain, F.S., et al., A performance based patient outcome score for active patients
following total knee arthroplasty. Osteoarthritis Cartilage, 2013. 21(1): p. 51-9.
203
209. Noble, P.C., et al., Development of a new Knee Society scoring system. Clin Orthop Relat
Res, 2012. 470(1): p. 20-32.
210. Bourne, R.B., et al., Patient Satisfaction after Total Knee Arthroplasty Who is Satisfied
and Who is Not? Clinical Orthopaedics and Related Research, 2010. 468(1): p. 57-63.
211. Bullens, P.H.J., et al., Patient satisfaction after total knee arthroplasty - A comparison
between subjective and objective outcome assessments. Journal of Arthroplasty, 2001.
16(6): p. 740-747.
212. Dahm, D.L., et al., Patient-reported activity level after total knee arthroplasty. Journal of
Arthroplasty, 2008. 23(3): p. 401-407.
213. Giesinger, J.M., et al., Association of psychological status and patient-reported physical
outcome measures in joint arthroplasty: a lack of divergent validity. Health Qual Life
Outcomes, 2013. 11: p. 64.
214. Tracey, J., Inside the clinic: health professionals' role in their clients' psychological
rehabilitation. J Sport Rehabil, 2008. 17(4): p. 413-31.
215. Clement, D., M.D. Granquist, and M.M. Arvinen-Barrow, Psychosocial aspects of
athletic injuries as perceived by athletic trainers. J Athl Train, 2013. 48(4): p. 512-21.
216. Mahomed, N., et al., The self-administered patient satisfaction scale for primary hip and
knee arthroplasty. Arthritis, 2011. 2011: p. 591253.
217. Jones, C.A. and S. Pohar, Health-related quality of life after total joint arthroplasty: a
scoping review. Clin Geriatr Med, 2012. 28(3): p. 395-429.
218. Engh, G.A., N.L. Parks, and C.E. Whitney, A prospective randomized study of
bicompartmental vs. total knee arthroplasty with functional testing and short term
outcome. J Arthroplasty, 2014. 29(9): p. 1790-4.
219. Vissers, M.M., et al., Functional capacity and actual daily activity do not contribute to
patient satisfaction after total knee arthroplasty. BMC Musculoskelet Disord, 2010. 11:
p. 121.
220. Von Keudell, A., et al., Patient satisfaction after primary total and unicompartmental
knee arthroplasty: an age-dependent analysis. Knee, 2014. 21(1): p. 180-4.
221. Lau, R.L., et al., Patient satisfaction after total knee and hip arthroplasty. Clin Geriatr
Med, 2012. 28(3): p. 349-65.
222. Chan, E.Y., et al., Acute postoperative pain following hospital discharge after total knee
arthroplasty. Osteoarthritis Cartilage, 2013. 21(9): p. 1257-63.
223. Kim, T.K., et al., Functional disabilities and satisfaction after total knee arthroplasty in
female Asian patients. J Arthroplasty, 2010. 25(3): p. 458-464 e1-2.
224. Kim, S.J., et al., Patients still wish for key improvements after total knee arthroplasty.
Knee Surg Relat Res, 2015. 27(1): p. 24-33.
225. Nyland, J., et al., Total knee arthroplasty in motivated patients with knee osteoarthritis
and athletic activity approach type goals: a conceptual decision-making model. Disabil
Rehabil, 2011. 33(17-18): p. 1683-92.
226. Devers, B.N., et al., Does greater knee flexion increase patient function and satisfaction
after total knee arthroplasty? J Arthroplasty, 2011. 26(2): p. 178-86.
227. Christen, M., E. Aghayev, and B. Christen, Short-term functional versus patient-reported
outcome of the bicruciate stabilized total knee arthroplasty: prospective consecutive case
series. BMC Musculoskelet Disord, 2014. 15: p. 435.
228. Mannion, A.F., et al., The role of patient expectations in predicting outcome after total
knee arthroplasty. Arthritis Res Ther, 2009. 11(5): p. R139.
204
229. Pua, Y.H., et al., Development of a Prediction Model to Estimate the Risk of Walking
Limitations in Patients with Total Knee Arthroplasty. J Rheumatol, 2016. 43(2): p. 419-
26.
230. Hepinstall, M.S., et al., Factors that impact expectations before total knee arthroplasty. J
Arthroplasty, 2011. 26(6): p. 870-6.
231. Su, E.P., et al., A prospective, multi-center, randomised trial to evaluate the efficacy of a
cryopneumatic device on total knee arthroplasty recovery. J Bone Joint Surg Br, 2012.
94(11 Suppl A): p. 153-6.
232. Thomas, K.M. and K.A. Sethares, Is guided imagery effective in reducing pain and
anxiety in the postoperative total joint arthroplasty patient? Orthop Nurs, 2010. 29(6): p.
393-9.
233. Sayeed, S.A., et al., The Risk of Subsequent Joint Arthroplasty after Primary Unilateral
Total Knee Arthroplasty, a 10-Year Study. Journal of Arthroplasty, 2011. 26(6): p. 842-
846.
234. Llikavainio, T., et al., Loading and gait symmetry during level and stair walking in
asymptomatic subjects with knee osteoarthritis: Importance of quadriceps femoris in
reducing impact force during heel strike? Knee, 2007. 14(3): p. 231-238.
235. Wiik, A., Gait Analysis Using an Instrumented Treadmill, in Department of Surgery and
Cancer. 2016, Imperial College London. p. 129.
236. Robinson, R.O., W. Herzog, and B.M. Nigg, Use of Force Platform Variables to
Quantify the Effects of Chiropractic Manipulation on Gait Symmetry. Journal of
Manipulative and Physiological Therapeutics, 1987. 10(4): p. 172-176.
237. Shao, Y.C., et al., The Fate of the Remaining Knee(s) or Hip(s) in Osteoarthritic Patients
Undergoing a Primary TKA or THA. Journal of Arthroplasty, 2013. 28(10): p. 1842-
1845.
238. Pozzi, F., L. Snyder-Mackler, and J. Zeni, Relationship between biomechanical
asymmetries during a step up and over task and stair climbing after total knee
arthroplasty. Clinical Biomechanics, 2015. 30(1): p. 78-85.
239. Yu, B., et al., Reproducibility of the kinematics and kinetics of the lower extremity during
normal stair-climbing. Journal of Orthopaedic Research, 1997. 15(3): p. 348-352.
240. Bade, M.J., W.M. Kohrt, and J.E. Stevens-Lapsley, Outcomes before and after total knee
arthroplasty compared to healthy adults. J Orthop Sports Phys Ther, 2010. 40(9): p. 559-
67.
241. Furu, M., et al., Quadriceps strength affects patient satisfaction after total knee
arthroplasty. J Orthop Sci, 2016. 21(1): p. 38-43.
242. Ward, T.R., et al., Bicruciate-stabilised total knee replacements produce more normal
sagittal plane kinematics than posterior-stabilised designs. J Bone Joint Surg Br, 2011.
93(7): p. 907-13.
243. Ishii, Y., et al., Quadriceps strength impairment in the mid- to long-term follow-up period
after total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc, 2016.
244. Christiansen, C.L., et al., Weight-bearing asymmetry during sit-stand transitions related
to impairment and functional mobility after total knee arthroplasty. Arch Phys Med
Rehabil, 2011. 92(10): p. 1624-9.
245. Collados-Maestre, I., et al., Better functional outcome after single-radius TKA compared
with multi-radius TKA. Knee Surg Sports Traumatol Arthrosc, 2016.
205
246. Morra, E.A., et al., The influence of contemporary knee design on high flexion: a
kinematic comparison with the normal knee. J Bone Joint Surg Am, 2008. 90 Suppl 4: p.
195-201.
247. Victor, J., et al., In vivo kinematics after a cruciate-substituting TKA. Clin Orthop Relat
Res, 2010. 468(3): p. 807-14.
248. Ebert, J.R., C. Munsie, and B. Joss, Guidelines for the early restoration of active knee
flexion after total knee arthroplasty: implications for rehabilitation and early
intervention. Arch Phys Med Rehabil, 2014. 95(6): p. 1135-40.
249. Tolk, J.J., et al., Total Knee Arthroplasty: What to Expect? A Survey of the Members of
the Dutch Knee Society on Long-Term Recovery after Total Knee Arthroplasty. J Knee
Surg, 2016.
250. Williams, D.H., D.S. Garbuz, and B.A. Marsi, Total knee arthroplasty: Techniques and
results. BC Medical Journal, 2010. 52(9): p. 447-454.
251. Bierke, S., M. Haner, and W. Petersen, Influence of somatization and depressive
symptoms on the course of pain within the first year after uncomplicated total knee
replacement: a prospective study. Int Orthop, 2016. 40(7): p. 1353-60.
252. Huijbregts, H.J., et al., Prosthetic alignment after total knee replacement is not
associated with dissatisfaction or change in Oxford Knee Score: A multivariable
regression analysis. Knee, 2016. 23(3): p. 535-9.
253. Akaike, H., New Look at Statistical-Model Identification. Ieee Transactions on Automatic
Control, 1974. Ac19(6): p. 716-723.
254. Lundblad, H., A. Kreicbergs, and K.A. Jansson, Prediction of persistent pain after total
knee replacement for osteoarthritis. Journal of Bone and Joint Surgery-British Volume,
2008. 90b(2): p. 166-171.
206
APPENDICES
207
Appendix A: Forgotten Joint Score
208
Appendix B: Informed Consents
Informed Consent Form for TKR Subjects
INFORMED CONSENT FORM Influence of Patient Satisfaction of Total knee Replacement Patients on Gait Mechanics during
Stair Negotiation and Walking
Principal Investigators: Songning Zhang, PhD. Kevin Valenzuela
Address: 340 HPER Address: 136 HPER
1914 Andy Holt Avenue 1914 Andy Holt Avenue
Knoxville, TN 37996 Knoxville, TN 37996
Phone: (865) 974-4716 Phone: (865) 974-2091
Introduction
You are invited to participate in this research study because you had a total knee
replacement (TKR) and are between 50 and 75 years old. The purpose of this research is to learn
the differences in how the knee works in people with total knee replacements (TKR). Before
agreeing to participate in this study, please read this form. It will tell you about the research.
Please ask us to explain anything you do not understand.
Testing Protocol
If you agree to be in this study, you will attend two study visits at the
Biomechanics/Sports Medicine Lab on the UT campus. Each visit will last about 1.5 – 2
hours. You can park on campus for free. You will fill out a few surveys (WOMAC, Forgotten
Joint Score, and Knee Society Score) about your knee and any problems you have. You will
need to wear shorts and t-shirt for the study procedures. Your shorts should be close-fitting so
we can see how your body moves during the study procedures. If you do not have close-fitting
shorts, we will provide paper laboratory shorts.
On the first visit, we will measure your weight and height. You will walk on the treadmill
for 3 minutes to get ready for the exercises. At different times during the study visit, we will ask
about your knee pain.
For the next tests, reflective markers will be placed on both sides of your feet, ankles, legs,
knees, thighs, pelvis and trunk. This is so a motion camera can follow your movements. You
will walk across the floor 3-5 times and climb up and down stairs 3-5 times. Each time you will
be asked to perform these movements at different speeds. You can take a rest between each
time. You can stop at any time. Data collected during the phone interview will also be used for
this study.
On the second day, you will do some exercises such as:
get out of a chair, walk about 9 feet, and walk back to the chair),
walk up and down a flight of stairs,
walk for 6 minutes,
209
stand up and sit down 10 times.
We will also perform some tests. Some of these tests will measure the activity of your
muscle. This is called electromyography (EMG). For those tests we will place electrodes on
several muscles. Electrodes record the electrical signal produced by the muscles. The electrodes
will not shock or hurt you. You will perform:
balance test,
strength test,
knee range of motion test.
Potential Risks
Risks associated with this study are minimal so they are no greater than your daily
activities. You can practice the movements to familiarize yourself with walking up and down the
staircase before the testing. You may stop anytime if you feel uncomfortable. If an injury
occurs during the course of study visit, we will provide standard first aid as necessary. In the
unlikely event you are injured during this study, the University of Tennessee does not
automatically provide reimbursement for medical care or other compensation. You will be
responsible for any medical expenses. If you are injured during the study, or you have questions,
please talk to Kevin Valenzuela or Songning Zhang (974-2091).
Every research study involves some risk to your confidentiality. It is possible that other
people could find out you were in the study or see your study information. But we will do our
best to keep your information confidential, so we think this risk is very low.
Benefits of Participation
You may not benefit from your participation in this study. If you want, you can receive
your individual study information to share with your personal physician in case it might be
helpful to your future health care.
Results from this study may help improve future designs, surgical methods and
rehabilitation used for TKR. Results may also help us understand why patients may be
dissatisfied with TKR and how patient satisfaction might be improved. You will also receive a
$60 gift card at the completion of the second testing session.
Confidentiality
Your information will be kept confidential. All research data and records will be stored
securely and will be made available only to researchers who work on this study. The motion
cameras will not record images of you. Your name will not be on any research data. Instead, a
code number will be assigned to all of your data.
Your name will not appear with the study results that will be presented at conferences
and published in journals. The research data will be kept indefinitely. These data may be used
210
for future research purposes after the completion of this project. If you withdraw from the study,
we will keep your consent form, but all of your other information will be destroyed.
Contact Information
If you have any questions about the study at any time or if you experience problems as a
result of participating in this study, you can contact Kevin Valenzuela or Songning Zhang at
1914 Andy Holt Ave. 136 HPER Bldg, The University of Tennessee (974-2091). Questions
about your rights as a research participant can be addressed to Compliance Officer in the Office
of Research at the University of Tennessee at (865) 974-7697.
Voluntary Participation and Withdrawal
Your participation is voluntary. You can refuse to participate and there will be no
penalty or loss of benefits to which you are otherwise entitled. You may withdraw from the study
at any time without penalty or loss of benefits to which you are otherwise entitled. Your
participation in this study may be stopped if you fail to follow the study procedures or if the
investigators feels that it is in your best interest to stop participation.
Consent Statement
I have read the above information. I agree to participate in this study. I have received a copy of
this form.
Subject’s Name: ___________________ Subject’s Signature: _____________ Date: _________
Investigator’s Signature: ____________________________ Date: __________
211
Informed Consent Form for Healthy Subjects
INFORMED CONSENT FORM Influence of Patient Satisfaction of Total knee Replacement Patients on Gait Mechanics during
Stair Negotiation and Walking
Principal Investigators: Songning Zhang, PhD. Kevin Valenzuela
Address: 340 HPER Address: 136 HPER
1914 Andy Holt Avenue 1914 Andy Holt Avenue
Knoxville, TN 37996 Knoxville, TN 37996
Phone: (865) 974-4716 Phone: (865) 974-2091
Introduction
You are invited to participate in this research study because you are a healthy adult and
are between 50 and 75 years old. The purpose of this research is to learn the differences in how
the knee works in people with total knee replacements (TKR) and those who are healthy. Before
agreeing to participate in this study, please read this form. It will tell you about the research.
Please ask us to explain anything you do not understand.
Testing Protocol
If you agree to be in this study, you will attend two study visits at the
Biomechanics/Sports Medicine Lab on the UT campus. Each visit will last about 1.5 – 2
hours. You can park on campus for free. You will fill out a few surveys (WOMAC, Forgotten
Joint Score, and Knee Society Score) about your knee and any problems you have. You will
need to wear shorts and t-shirt for the study procedures. Your shorts should be close-fitting so
we can see how your body moves during the study procedures. If you do not have close-fitting
shorts, we will provide paper laboratory shorts.
On the first day, we will measure your weight and height. You will walk on the treadmill
for 3 minutes to get ready for the exercises. At different times during the study visit, we will ask
about your knee pain.
For the next tests, reflective markers will be placed on both sides of your feet, ankles,
legs, knees, thighs, pelvis and trunk. This is so a motion camera can follow your
movements. You will walk across the floor 3-5 times and climb up and down stairs 3-5 times.
Each time you will be asked to perform these movements at different speeds. You can take a rest
between each time. You can stop at any time.
On the second day, you will do some exercises such as:
get out of a chair, walk about 9 feet, and walk back to the chair),
walk up and down a flight of stairs,
walk for 6 minutes,
212
stand up and sit down 10 times.
We will also perform some tests. Some of these tests will measure the activity of your
muscle. This is called electromyography (EMG). For those tests we will place electrodes on
several muscles. Electrodes record the electrical signal produced by the muscles. The electrodes
will not shock or hurt you. You will perform:
balance test,
strength test,
knee range of motion test.
Potential Risks
Risks associated with this study are minimal so they are no greater than your daily
activities. You can practice the movements to familiarize yourself with walking up and down the
staircase before the testing. You may stop anytime if you feel uncomfortable. If an injury
occurs during the course of study visit, we will provide standard first aid as necessary. In the
unlikely event your are injured during this study, the University of Tennessee does not
automatically provide reimbursement for medical care or other compensation. You will be
responsible for any medical expenses. If you are injured during the study, or you have questions,
please talk to Kevin Valenzuela or Songning Zhang (974-2091).
Every research study involves some risk to your confidentiality. It is possible that other
people could find out you were in the study or see your study information. But we will do our
best to keep your information confidential, so we think this risk is very low.
Benefits of Participation
You may not benefit from your participation in this study. If you want, you can receive
your individual study information to share with your personal physician in case it might be
helpful to your future health care.
Results from this study may help improve future designs, surgical methods and
rehabilitation used for TKR. Results may also help us understand why patients may be
dissatisfied with TKR and how patient satisfaction might be improved. Even though you do not
have a TKR, your participation may help to further the understanding of why TKR patients have
physical issues. You will also receive a $40 gift card at the completion of the second testing
session.
Confidentiality
Your information will be kept confidential. All research data and records will be stored
securely and will be made available only to researchers who work on this study. The motion
cameras will not record images of you. Your name will not be on any research data. Instead, a
code number will be assigned to all of your data.
213
Your name will not appear with the study results that will be presented at conferences
and published in journals. The research data will be kept indefinitely. These data may be used
for future research purposes after the completion of this project. If you withdraw from the study,
we will keep your consent form, but all of your other information will be destroyed.
Contact Information
If you have any questions about the study at any time or if you experience problems as a
result of participating in this study, you can contact Kevin Valenzuela or Songning Zhang at
1914 Andy Holt Ave. 136 HPER Bldg, The University of Tennessee (974-2091). Questions
about your rights as a research participant can be addressed to Compliance Officer in the Office
of Research at the University of Tennessee at (865) 974-7697.
Voluntary Participation and Withdrawal
Your participation is voluntary. You can refuse to participate and there will be no
penalty or loss of benefits to which you are otherwise entitled. You may withdraw from the study
at any time without penalty or loss of benefits to which you are otherwise entitled. Your
participation in this study may be stopped if you fail to follow the study procedures or if the
investigators feels that it is in your best interest to stop participation.
Consent Statement
I have read the above information. I agree to participate in this study. I have received a
copy of this form.
Subject’s Name: ___________________ Subject’s Signature: _____________ Date: _________
Investigator’s Signature: ____________________________ Date: __________
214
Appendix C: Physical Activity Readiness Questionnaire
215
Appendix D: Visual Analogue Pain Scale
216
Appendix E: Demographic Questionnaire
Information Sheet
Subject # ____________________________ Date (MM/DD/YY): _____/_____/_______
DOB (MM/DD/YY): _____/_____/_______ Shoe Size (US) _______________
Height: ___ Feet ___ Inches or ______ cm Weight: _________lbs or _________ kg
BMI (answered by PI) ______ Gender (circle one): Female Male
For knee replacement subjects
Knee replacement side (circle one): Right Left
Date of replacement (MM/YY): _____/_______
Knee replacement type (circle one):
Posterior Stabilized Mobile Bearing Fixed Bearing Other
If select “Other”, please provide more details below if known.
_________________________________________________________________________
Office _________________and doctor who performed surgery _____________________
Rehabilitation clinic and duration: ____________________________
How satisfied are you with your total knee replacement (Circle one)?
Very Dissatisfied Dissatisfied Neutral Satisfied Very Satisfied
Have you had any additional osteoarthritis, systemic arthritis, joint pathologies, or surgeries?
______________________________________________________________________________
Are you able to walk without a walking aid? ___________ Pregnant/Nursing? __________
Neurological conditions? _________________________________________________________
Are you currently taking any pain medications? Frequency? _____________________________
Any reason you should not be participating in exercise? ________________________________
217
Appendix F: Recruitment Flyers
Total Knee Replacement Recruitment Flyer
218
Healthy Subject Recruitment Flyer
219
Appendix G: Subject Demographics
Table 18. Dissatisfied TKR patient characteristics
Subject Gender Height (m) Weight (kg) BMI (kg/m2) Age (years) Replaced limb Time from surgery (months)
S4 F 1.70 78.9 27.30 59 R 25
S6 M 1.76 112.28 36.25 70 L 25
S13 F 1.71 73.7 25.35 66 R 24
S18 M 1.71 74.46 25.61 72 L 20
S20 F 1.67 85.83 30.78 72 R 24
S28 F 1.70 75.59 26.31 67 L 34
S30 F 1.52 51.27 22.19 65 L 53
S32 M 1.77 106.32 34.13 71 L 55
S35 F 1.67 70.54 25.29 70 R 51
Mean - 1.69 80.99 28.13 68.0 - 34.5
220
Table 19. Satisfied TKR patient characteristics
Subject Gender Height (m) Weight (kg) BMI (kg/m2) Age (years) Replaced limb Time from surgery (months)
S1 M 1.91 108.26 29.68 73 L 17
S2 M 1.875 114.17 32.48 54 L 21
S3 F 1.68 85.42 30.27 57 R 27
S7 F 1.65 62.39 22.92 60 R 20
S8 M 1.88 106.32 30.08 67 L 40
S9 M 1.795 102.55 31.83 64 R 29
S10 F 1.675 63.4 22.60 68 R 54
S11 M 1.68 107.07 37.94 72 R 12
S12 F 1.715 91.95 31.26 64 R 26
S14 M 1.76 79.85 25.78 75 R 53
S16 M 1.805 92.71 28.46 67 R 45
S17 F 1.63 80.22 30.19 70 L 27
S19 M 1.85 79.61 23.26 74 L 20
S29 M 1.89 104.59 29.28 71 R 22
S36 M 1.67 72.38 25.95 63 R 26
Mean - 1.76 90.19 28.84 66.6 - 29.3
221
Table 20. Healthy control participant characteristics
Subject Gender Height (m) Weight (kg) BMI (kg/m2) Age (years) Dominant limb
S5 F 1.805 81.86 25.13 53 R
S15 M 1.905 79.2 21.82 59 R
S21 M 1.72 83.38 28.18 52 R
S22 F 1.72 58.51 19.78 55 R
S23 M 1.91 75.54 20.71 73 R
S24 F 1.665 66.56 24.01 57 R
S25 M 1.795 94.5 29.33 50 R
S26 F 1.685 71.86 25.31 57 L
S27 M 1.705 83.03 28.56 54 R
S31 F 1.62 63.81 24.31 62 R
S33 F 1.76 99.18 32.02 50 R
S34 M 1.855 92.97 27.02 70 R
S37 F 1.755 77.37 25.12 71 R
S38 M 1.76 70.85 22.87 75 R
S39 F 1.62 67.53 25.73 73 R
Mean - 1.75 77.74 25.32 60.7 -
222
Appendix H: Individual Results for Select Variables
Table 21. WOMAC subscales and total scores and passive knee ROM (°) for Dissatisfied TKR patients. Subject Limb Pain Stiffness Physical Function Total Passive ROM
S4 Replaced 471 172 908.5 1551.5 122
S6 Replaced 169 7 647 823 113
S13 Replaced 45.5 76 100 221.5 112
S18 Replaced 128 96.5 278.5 503 116
S20 Replaced 292.5 154 765 1211.5 122
S28 Replaced 282 91 815 1188 105
S30 Replaced 59 29 89.5 177.5 127
S32 Replaced 77.5 32.5 335.5 445.5 121
S35 Replaced 218.5 27 787.5 1033 127
S4 Non-Replaced 10 5 34.5 49.5 146
S6 Non-Replaced 35 10 53.5 98.5 112
S13 Non-Replaced 13.5 6 53.5 73 130
S18 Non-Replaced 0 0 0 0 135
S20 Non-Replaced 7.5 3 32 42.5 127
S28 Non-Replaced 106.5 16 91 213.5 132
S30 Non-Replaced 14 7 63.5 84.5 137
S32 Non-Replaced 0 0 0 0 120
S35 Non-Replaced 0 27 16.5 43.5 140
Mean Replaced 193.7 76.1 525.2 794.7 118.3
Mean Non-Replaced 20.7 8.2 38.3 67.2 131.0
223
Table 22. WOMAC subscales and total scores and passive knee ROM (°) for Satisfied TKR patients. Subject Limb Pain Stiffness Physical Function Total Passive ROM
S1 Replaced 63 68 543 674 132
S2 Replaced 17 16 84 117 125
S3 Replaced 18 4.5 99.5 122 125
S7 Replaced 16.5 22 147.5 186 135
S8 Replaced 6.5 150 187 343.5 105
S9 Replaced 16.5 9 78.5 104 124
S10 Replaced 12 8 89.5 109.5 110
S11 Replaced 69.5 6 252.5 328 120
S12 Replaced 20 91.5 69 180.5 107
S14 Replaced 14.5 6.5 51 72 131
S16 Replaced 108 68.5 331.5 508 116
S17 Replaced 6 2 104.5 112.5 130
S19 Replaced 69 27 225 321 140
S29 Replaced 28 53 363.5 444.5 129
S36 Replaced 44 39.5 62 145.5 115
S1 Non-Replaced 47 38 403.5 488.5 150
S2 Non-Replaced 8 11.5 62 81.5 135
S3 Non-Replaced 11 4.5 49.5 65 134
S7 Non-Replaced 12.5 13.5 66 92 141
S8 Non-Replaced 6.5 150 187 343.5 113
S9 Non-Replaced 29 15.5 131.5 176 132
S10 Non-Replaced 12 8 50.5 70.5 136
S11 Non-Replaced 69.5 6 197.5 273 127
S12 Non-Replaced 6 0 0 6 121
S14 Non-Replaced 14.5 6.5 51 72 139
S16 Non-Replaced 7.5 3 42 52.5 131
S17 Non-Replaced 6 2 41 49 135
S19 Non-Replaced 69 27 225 321 147
S29 Non-Replaced 48.5 114 247 409.5 134
S36 Non-Replaced 30 3 0 33 141
Mean Replaced 33.9 38.1 179.2 251.2 122.9
Mean Non-Replaced 25.1 26.8 144.5 196.5 134.4
224
Table 23. WOMAC subscales and total scores and passive knee ROM (°) for healthy controls. Subject Limb Pain Stiffness Physical Function Total Passive ROM
S5 Dominant 62 69 129.5 260.5 95
S15 Dominant 27 23.5 92.5 143 139
S21 Dominant 0 0 0 0 139
S22 Dominant 0 0 0 0 153
S23 Dominant 0 0 0 0 130
S24 Dominant 0 0 0 0 138
S25 Dominant 20 0 0 20 125
S26 Dominant 0 0 0 0 140
S27 Dominant 0 0 0 0 145
S31 Dominant 0 0 0 0 130
S33 Dominant 0 5.5 11 16.5 132
S34 Dominant 0 0 9 9 124
S37 Dominant 0 0 0 0 139
S38 Dominant 0 0 0 0 148
S39 Dominant 0 0 0 0 137
S5 Non- Dominant 13.5 6 58 77.5 96
S15 Non- Dominant 27 23.5 92.5 143 139
S21 Non- Dominant 0 0 0 0 140
S22 Non- Dominant 0 0 0 0 150
S23 Non- Dominant 0 0 0 0 128
S24 Non- Dominant 0 0 0 0 135
S25 Non- Dominant 0 0 0 0 125
S26 Non- Dominant 6.5 8 17 31.5 145
S27 Non- Dominant 0 0 0 0 147
S31 Non- Dominant 0 0 0 0 140
S33 Non- Dominant 0 5.5 11 16.5 137
S34 Non- Dominant 0 0 9 9 125
S37 Non- Dominant 0 0 0 0 143
S38 Non- Dominant 0 0 0 0 149
S39 Non- Dominant 0 0 0 0 137
Mean Dominant 7.3 6.5 16.1 29.9 134.3
Mean Non- Dominant 3.1 2.9 12.5 18.5 135.4
225
Table 24. Functional testing times and over ground walking, stair ascent, and stair descent velocity for controls and TKR patients. Subject Group Chair Rise (s) Stair Ascend (s) Stair Descend (s) Walking (m/s) Ascent (m/s) Descent (m/s)
S4 Dissatisfied 25.11 4.74 4.3 1.35 0.74 0.67
S6 Dissatisfied 11.56 4.47 4.41 1.07 0.49 0.40
S13 Dissatisfied 16.42 4.37 4.04 1.50 0.77 0.59
S18 Dissatisfied 28.99 4.18 4.04 1.13 0.54 0.43
S20 Dissatisfied 18.84 4.84 4.27 1.10 0.51 0.44
S28 Dissatisfied 20.65 6.58 7.71 0.99 0.40 0.37
S30 Dissatisfied 10.01 4.67 4.6 1.18 0.56 0.64
S32 Dissatisfied 12.63 5.42 7.23 1.05 0.48 0.34
S35 Dissatisfied 15.42 10.27 12.74 1.00 0.21 0.15
S5 Healthy 7.26 3.99 3.61 1.39 0.98 0.70
S15 Healthy 13.92 3.28 2.47 1.38 0.67 0.76
S21 Healthy 8.97 3.04 3.53 1.24 0.61 0.54
S22 Healthy 12.89 3.56 3.38 1.60 0.89 0.78
S23 Healthy 21.02 4.29 3.84 1.23 0.58 0.56
S24 Healthy 20.26 2.96 3.38 1.50 0.76 0.63
S25 Healthy 16.8 4.31 3.17 1.31 0.70 0.70
S26 Healthy 16.25 4.16 3.63 1.47 0.77 0.71
S27 Healthy 14.99 4.81 4.2 1.34 0.62 0.63
S31 Healthy 15.29 4.37 4.13 1.13 0.52 0.50
S33 Healthy 24.12 4.61 3.77 1.25 0.68 0.64
S34 Healthy 29.89 4.27 3.69 1.22 0.63 0.66
S37 Healthy 15.66 4.16 4.12 1.36 0.72 0.67
S38 Healthy 18.83 3.58 3.69 1.35 0.64 0.56
S39 Healthy 19.97 5.49 4.59 1.19 0.57 0.50
S1 Satisfied 16.14 5.54 5.05 1.45 0.94 0.75
S2 Satisfied 17.3 4.97 3.83 1.37 0.72 0.73
S3 Satisfied 12.5 4.66 3.87 1.48 0.87 0.78
S7 Satisfied 13.94 4.21 4.08 1.30 0.60 0.57
S8 Satisfied 9.91 5.28 4.83 1.28 0.69 0.49
S9 Satisfied 10.8 3.47 3.29 1.33 0.82 0.66
S10 Satisfied 12.63 4.1 3.23 1.38 0.76 0.70
S11 Satisfied 13 4.27 4.17 1.12 0.60 0.52
S12 Satisfied 22.35 4.55 4.39 1.21 0.68 0.54
S14 Satisfied 15.59 3.86 3.74 1.40 0.78 0.68
S16 Satisfied 25.11 2.94 2.98 1.32 0.70 0.59
S17 Satisfied 11.56 4.18 3.98 1.20 0.76 0.71
S19 Satisfied 16.42 3.38 4.14 1.18 0.82 0.66
S29 Satisfied 28.99 5.48 4.7 1.30 0.78 0.61
S36 Satisfied 18.84 3.58 3.4 1.43 0.64 0.59
Mean Dissatisfied 18.43 5.50 5.93 1.15 0.52 0.45
Mean Satisfied 16.84 4.30 3.98 1.32 0.74 0.64
Mean Healthy 15.01 4.06 3.68 1.33 0.69 0.63
226
Table 25. 1st and 2nd peak VGRF for Dissatisfied TKR patients. 1st peak VGRF (BW) 2nd peak VGRF (BW)
Subject Limb Walking Ascent Descent Walking Ascent Descent
S4 Replaced 1.185±0.023 0.962±0.012 1.583±0.133 1.099±0.014 1.203±0.034 0.867±0.060
S6 Replaced 1.015±0.013 0.883±0.026 1.118±0.101 0.984±0.033 1.035±0.037 0.947±0.012
S13 Replaced 1.033±0.020 0.969±0.006 1.359±0.052 1.056±0.021 0.993±0.036 0.885±0.020
S18 Replaced 1.093±0.031 0.968±0.042 1.064±0.079 0.991±0.016 1.066±0.045 0.890±0.041
S20 Replaced 0.998±0.017 0.809±0.122 1.216±0.070 0.946±0.013 1.049±0.029 0.953±0.021
S28 Replaced 0.978±0.030 0.816±0.014 1.187±0.094 0.998±0.011 0.907±0.014 0.927±0.037
S30 Replaced 1.045±0.036 1.026±0.027 1.680±0.329 1.018±0.013 1.098±0.068 0.839±0.052
S32 Replaced 0.984±0.013 0.937±0.024 1.023±0.052 0.972±0.012 0.995±0.032 0.998±0.030
S35 Replaced 0.946±0.022 0.971±0.057 0.950±0.029 1.002±0.011 0.951±0.084 0.912±0.040
S4 Non-Replaced 1.133±0.057 0.949±0.024 1.345±0.098 1.089±0.029 1.156±0.038 0.771±0.034
S6 Non-Replaced 1.037±0.023 0.999±0.013 1.476±0.061 1.058±0.012 1.106±0.022 0.912±0.027
S13 Non-Replaced 1.179±0.022 1.036±0.011 1.415±0.038 1.132±0.030 1.147±0.025 1.037±0.037
S18 Non-Replaced 1.113±0.040 1.024±0.024 1.026±0.090 1.044±0.031 1.166±0.023 1.088±0.041
S20 Non-Replaced 1.056±0.021 0.938±0.022 1.461±0.156 1.022±0.013 1.219±0.137 0.960±0.029
S28 Non-Replaced 1.005±0.012 1.014±0.041 1.402±0.055 0.992±0.020 0.978±0.010 0.993±0.055
S30 Non-Replaced 1.178±0.056 1.074±0.022 1.613±0.166 1.046±0.012 1.211±0.050 0.939±0.029
S32 Non-Replaced 1.007±0.012 1.018±0.026 1.116±0.063 0.985±0.012 1.121±0.054 0.990±0.022
S35 Non-Replaced 1.053±0.039 1.033±0.022 1.236±0.109 1.003±0.013 0.992±0.031 1.034±0.013
Mean Replaced 1.03±0.07 0.93±0.07 1.24±0.25 1.01±0.05 1.03±0.09 0.91±0.05
Mean Non-Replaced 1.08±0.07 1.01±0.04 1.34±0.19 1.04±0.05 1.12±0.09 0.97±0.09
227
Table 26. 1st and 2nd peak VGRF for Satisfied TKR patients. 1st peak VGRF (BW) 2nd peak VGRF (BW)
Subject Limb Walking Ascent Descent Walking Ascent Descent
S1 Replaced 1.100±0.019 0.366±0.022 1.235±0.092 1.004±0.027 1.164±0.051 0.878±0.033
S2 Replaced 1.079±0.017 0.354±0.024 1.322±0.087 1.080±0.010 1.022±0.095 0.861±0.061
S3 Replaced 1.188±0.021 0.357±0.036 1.701±0.066 1.056±0.015 1.247±0.039 0.848±0.076
S7 Replaced 0.985±0.018 0.394±0.021 1.280±0.118 1.050±0.013 1.103±0.037 0.881±0.031
S8 Replaced 0.990±0.012 0.483±0.018 1.401±0.091 1.080±0.034 1.164±0.057 0.976±0.028
S9 Replaced 1.116±0.021 0.404±0.024 1.619±0.088 1.109±0.026 1.186±0.053 0.905±0.031
S10 Replaced 1.035±0.027 0.380±0.008 1.810±0.066 1.096±0.025 1.106±0.061 0.777±0.013
S11 Replaced 1.033±0.031 0.497±0.043 1.640±0.034 0.973±0.024 1.182±0.061 0.887±0.051
S12 Replaced 1.042±0.010 0.419±0.030 1.534±0.056 1.006±0.015 1.165±0.015 0.907±0.031
S14 Replaced 1.146±0.057 0.347±0.018 1.692±0.123 1.105±0.029 1.274±0.030 0.918±0.086
S16 Replaced 1.140±0.028 0.459±0.005 1.404±0.179 1.087±0.039 1.216±0.074 0.898±0.039
S17 Replaced 1.073±0.059 0.365±0.035 1.548±0.105 1.117±0.022 1.122±0.019 0.807±0.043
S19 Replaced 1.033±0.005 0.378±0.011 1.747±0.178 1.017±0.018 1.156±0.041 0.811±0.008
S29 Replaced 1.108±0.012 0.400±0.019 1.556±0.086 1.070±0.009 1.064±0.034 0.816±0.022
S36 Replaced 1.149±0.022 0.422±0.013 1.270±0.086 1.150±0.016 1.201±0.057 0.923±0.018
S1 Non-Replaced 1.259±0.041 1.078±0.050 1.637±0.060 1.076±0.023 1.245±0.031 0.931±0.076
S2 Non-Replaced 1.081±0.032 1.044±0.029 1.378±0.072 1.088±0.009 1.115±0.048 0.941±0.034
S3 Non-Replaced 1.312±0.021 0.935±0.016 1.963±0.071 1.156±0.030 1.210±0.039 0.843±0.105
S7 Non-Replaced 1.145±0.046 1.043±0.019 1.821±0.197 1.124±0.029 1.366±0.078 1.076±0.030
S8 Non-Replaced 1.015±0.024 0.961±0.009 1.319±0.064 1.060±0.007 1.093±0.018 0.912±0.012
S9 Non-Replaced 1.105±0.024 1.026±0.020 1.550±0.051 1.113±0.015 1.105±0.014 0.912±0.037
S10 Non-Replaced 1.104±0.021 1.071±0.038 1.917±0.060 1.143±0.031 1.161±0.060 0.907±0.027
S11 Non-Replaced 1.106±0.021 0.952±0.005 1.658±0.088 0.997±0.019 1.156±0.072 0.913±0.030
S12 Non-Replaced 1.078±0.034 0.997±0.014 1.462±0.065 1.007±0.007 1.248±0.031 0.969±0.049
S14 Non-Replaced 1.094±0.017 1.041±0.021 1.660±0.061 1.141±0.033 1.316±0.015 0.868±0.021
S16 Non-Replaced 1.144±0.038 1.054±0.034 1.599±0.133 1.099±0.011 1.389±0.054 1.028±0.054
S17 Non-Replaced 1.078±0.028 1.063±0.011 1.410±0.045 1.076±0.033 1.162±0.056 0.824±0.083
S19 Non-Replaced 1.065±0.030 0.812±0.268 1.611±0.147 1.055±0.010 1.368±0.264 0.959±0.032
S29 Non-Replaced 1.135±0.050 1.012±0.027 1.406±0.080 1.073±0.023 1.074±0.022 0.887±0.052
S36 Non-Replaced 1.201±0.011 1.107±0.025 1.520±0.106 1.199±0.009 1.366±0.066 1.005±0.036
Mean Replaced 1.08±0.06 0.99±0.04 1.52±0.19 1.07±0.05 1.16±0.07 0.87±0.05
Mean Non-Replaced 1.13±0.08 1.01±0.07 1.59±0.19 1.09±0.05 1.22±0.11 0.93±0.07
228
Table 27. 1st and 2nd peak VGRF for healthy controls. 1st peak VGRF (BW) 2nd peak VGRF (BW)
Subject Limb Walking Ascent Descent Walking Ascent Descent
S5 Dominant 1.189±0.020 1.136±0.024 1.403±0.028 0.981±0.033 0.927±0.085 0.791±0.069
S15 Dominant 1.116±0.004 1.017±0.043 1.520±0.114 1.107±0.035 1.025±0.094 0.660±0.061
S21 Dominant 1.180±0.035 1.071±0.007 1.482±0.054 1.114±0.017 1.104±0.036 0.866±0.046
S22 Dominant 1.194±0.033 1.074±0.030 1.537±0.104 1.268±0.020 1.314±0.047 1.040±0.034
S23 Dominant 1.178±0.014 1.052±0.012 1.371±0.068 1.164±0.016 1.044±0.026 0.902±0.022
S24 Dominant 1.197±0.011 1.183±0.011 1.348±0.061 1.256±0.043 1.314±0.021 0.980±0.027
S25 Dominant 1.147±0.038 0.942±0.023 1.335±0.083 1.082±0.009 1.171±0.050 0.892±0.053
S26 Dominant 1.193±0.029 1.010±0.025 1.444±0.064 1.138±0.014 1.157±0.062 0.874±0.055
S27 Dominant 1.227±0.017 0.673±0.141 1.594±0.097 1.055±0.029 1.108±0.038 0.785±0.046
S31 Dominant 1.137±0.021 1.061±0.023 1.484±0.096 1.092±0.019 1.172±0.027 1.034±0.052
S33 Dominant 1.073±0.007 0.953±0.022 1.418±0.027 1.060±0.017 1.117±0.031 0.902±0.024
S34 Dominant 1.209±0.022 1.056±0.034 1.397±0.042 1.029±0.015 0.969±0.050 0.798±0.063
S37 Dominant 1.033±0.022 1.021±0.018 1.358±0.109 1.061±0.044 1.174±0.057 0.908±0.040
S38 Dominant 1.169±0.038 0.993±0.024 1.567±0.067 1.047±0.014 1.199±0.036 0.886±0.024
S39 Dominant 1.052±0.049 1.000±0.024 1.333±0.020 1.014±0.032 1.257±0.021 0.989±0.018
S5 Non- Dominant 1.182±0.033 1.206±0.070 1.505±0.051 1.100±0.006 1.084±0.061 0.769±0.026
S15 Non- Dominant 1.167±0.033 0.980±0.036 1.635±0.190 1.078±0.025 1.010±0.075 0.797±0.063
S21 Non- Dominant 1.129±0.026 0.997±0.022 1.379±0.069 1.074±0.027 1.018±0.029 0.876±0.017
S22 Non- Dominant 1.143±0.022 1.107±0.037 1.566±0.082 1.249±0.014 1.379±0.056 1.018±0.008
S23 Non- Dominant 1.155±0.029 1.101±0.017 1.429±0.054 1.106±0.026 1.052±0.027 0.943±0.016
S24 Non- Dominant 1.192±0.015 1.147±0.015 1.368±0.047 1.218±0.023 1.259±0.026 1.038±0.053
S25 Non- Dominant 1.100±0.021 0.942±0.037 1.376±0.042 1.075±0.026 1.195±0.078 0.942±0.021
S26 Non- Dominant 1.145±0.019 0.997±0.028 1.506±0.066 1.086±0.021 1.186±0.060 0.776±0.024
S27 Non- Dominant 1.280±0.011 1.122±0.018 1.345±0.027 1.118±0.028 1.204±0.055 0.779±0.021
S31 Non- Dominant 1.097±0.022 1.011±0.006 1.508±0.072 1.067±0.015 1.110±0.037 0.912±0.037
S33 Non- Dominant 1.054±0.016 0.942±0.018 1.366±0.026 1.015±0.015 1.059±0.068 0.847±0.014
S34 Non- Dominant 1.122±0.016 1.028±0.028 1.351±0.090 0.981±0.023 0.994±0.031 0.783±0.046
S37 Non- Dominant 1.093±0.011 0.998±0.032 1.385±0.092 1.092±0.016 1.222±0.011 0.902±0.054
S38 Non- Dominant 1.131±0.006 1.002±0.079 1.370±0.163 1.072±0.004 1.121±0.049 0.890±0.016
S39 Non- Dominant 1.064±0.023 1.028±0.024 1.304±0.047 1.038±0.008 1.186±0.049 0.988±0.034
Mean Dominant 1.15±0.06 1.02±0.11 1.44±0.09 1.10±0.08 1.14±0.11 0.89±0.10
Mean Non- Dominant 1.14±0.06 1.04±0.08 1.43±0.09 1.09±0.07 1.14±0.11 0.88±0.09
229
Table 28. Knee Flexion ROM (°) and loading-response knee extension moment (Nm/kg) for Dissatisfied TKR patients. Flexion ROM Loading-response Extension Moment
Subject Limb Walking Ascent Descent Walking Ascent Descent
S4 Replaced -9.990±1.196 57.396±1.954 -82.870±2.412 0.532±0.054 1.039±0.076 0.764±0.114
S6 Replaced -16.467±2.616 49.496±2.869 -79.736±0.807 0.496±0.062 0.815±0.068 0.320±0.096
S13 Replaced -1.744±0.788 57.186±1.646 -77.057±2.564 0.132±0.029 0.820±0.031 0.614±0.077
S18 Replaced -18.455±2.265 55.057±2.150 -82.720±2.443 0.691±0.069 1.187±0.086 0.324±0.169
S20 Replaced -14.268±1.847 44.907±2.711 -82.885±2.540 0.519±0.053 0.626±0.058 0.397±0.092
S28 Replaced -5.424±1.098 52.999±2.360 -83.714±1.243 0.265±0.090 0.086±0.015 0.776±0.019
S30 Replaced -17.827±1.675 52.421±3.130 -89.131±2.055 0.361±0.057 0.800±0.085 0.397±0.200
S32 Replaced -12.168±0.767 50.049±3.033 -78.646±2.128 0.348±0.063 0.819±0.043 0.047±0.073
S35 Replaced -3.144±0.817 61.311±2.134 -85.607±1.912 -0.034±0.017 0.538±0.031 0.588±0.078
S4 Non-Replaced -12.288±1.547 63.672±1.062 -83.566±1.612 0.719±0.017 1.184±0.061 1.004±0.057
S6 Non-Replaced -18.528±1.949 63.002±1.573 -80.337±1.896 0.824±0.067 1.422±0.009 0.834±0.071
S13 Non-Replaced -14.395±1.895 62.426±0.352 -80.284±0.514 1.155±0.074 1.335±0.057 1.133±0.149
S18 Non-Replaced -18.207±3.459 54.717±1.693 -82.690±1.741 0.613±0.054 1.319±0.047 0.296±0.159
S20 Non-Replaced -15.508±1.867 49.512±8.143 -81.620±2.315 0.496±0.032 0.956±0.053 0.608±0.143
S28 Non-Replaced -8.400±1.792 56.977±2.569 -87.574±1.230 0.069±0.052 0.591±0.120 0.357±0.111
S30 Non-Replaced -16.025±1.169 58.546±1.880 -89.691±2.623 0.509±0.101 1.147±0.068 0.908±0.214
S32 Non-Replaced -17.483±1.152 54.052±3.824 -81.210±1.280 0.212±0.058 0.890±0.032 0.176±0.070
S35 Non-Replaced -17.843±2.788 49.955±4.792 -81.460±0.506 0.486±0.081 1.152±0.100 0.853±0.181
Mean Replaced -11.1±6.4 53.4±4.9 -82.5±3.7 0.42±0.18 0.75±0.31 0.39±0.22
Mean Non-Replaced -15.4±3.3 57.0±5.4 -83.2±3.3 0.56±0.32 1.11±0.26 0.69±0.34
230
Table 29. Knee Flexion ROM (°) and loading-response knee extension moment (Nm/kg) for Satisfied TKR patients. Flexion ROM Loading-response Extension Moment
Subject Limb Walking Ascent Descent Walking Ascent Descent
S1 Replaced -16.688±2.491 57.161±1.304 -78.017±1.145 0.712±0.049 1.002±0.073 0.775±0.177
S2 Replaced -13.383±1.910 48.854±1.323 -72.199±3.429 0.516±0.031 1.486±0.052 0.993±0.209
S3 Replaced -22.634±2.745 57.280±2.674 -85.243±1.932 0.856±0.036 1.108±0.067 0.826±0.160
S7 Replaced -6.819±1.111 63.833±1.798 -90.844±1.168 0.112±0.014 0.888±0.039 0.629±0.078
S8 Replaced -17.432±1.169 51.666±1.658 -72.391±1.143 0.559±0.042 1.422±0.042 0.790±0.113
S9 Replaced -16.363±2.857 42.851±2.447 -79.359±0.260 0.538±0.068 1.203±0.031 0.925±0.171
S10 Replaced -14.097±1.338 52.127±1.001 -78.754±1.337 0.568±0.043 1.047±0.050 0.947±0.131
S11 Replaced -11.663±0.758 58.939±1.655 -76.861±1.658 0.606±0.028 0.976±0.028 0.859±0.072
S12 Replaced -6.046±4.071 57.114±1.135 -83.544±2.295 0.305±0.056 0.670±0.021 0.244±0.106
S14 Replaced -10.787±2.207 57.931±2.126 -75.023±1.996 0.460±0.056 1.220±0.047 0.912±0.192
S16 Replaced -13.879±1.874 62.448±8.253 -82.498±3.229 0.869±0.051 1.391±0.131 0.843±0.262
S17 Replaced -8.989±2.065 51.111±1.553 -69.572±1.615 0.635±0.091 1.114±0.058 0.906±0.116
S19 Replaced -3.295±0.697 57.233±2.755 -76.338±1.610 .±. 1.075±0.038 0.392±0.196
S29 Replaced -14.403±0.682 56.418±2.492 -76.624±2.128 0.507±0.045 1.009±0.037 0.610±0.076
S36 Replaced -13.370±2.162 54.514±2.553 -85.018±1.325 0.410±0.053 0.558±0.056 0.425±0.109
S1 Non-Replaced -20.404±1.368 52.054±0.830 -71.598±1.837 1.075±0.156 1.498±0.074 1.227±0.227
S2 Non-Replaced -11.042±1.471 53.177±1.719 -73.581±3.150 0.338±0.038 1.545±0.070 1.004±0.153
S3 Non-Replaced -25.190±2.345 60.936±3.053 -83.176±3.164 1.259±0.090 1.122±0.072 0.748±0.054
S7 Non-Replaced -21.408±0.695 48.002±2.963 -90.423±2.292 0.665±0.098 1.361±0.041 1.019±0.092
S8 Non-Replaced -12.573±1.516 59.403±0.920 -77.041±1.156 0.468±0.068 1.084±0.051 0.470±0.088
S9 Non-Replaced -23.817±0.605 59.178±1.808 -80.224±2.025 0.720±0.052 1.429±0.052 0.674±0.166
S10 Non-Replaced -17.877±1.040 59.595±3.278 -85.110±1.698 0.716±0.074 1.124±0.062 1.161±0.168
S11 Non-Replaced -18.452±1.089 53.137±1.287 -81.966±0.368 0.686±0.059 0.984±0.034 0.884±0.096
S12 Non-Replaced -14.515±0.468 53.759±1.579 -88.573±1.113 0.577±0.048 1.303±0.033 1.077±0.142
S14 Non-Replaced -16.745±1.781 65.652±1.764 -82.736±1.187 0.520±0.069 1.146±0.079 0.359±0.352
S16 Non-Replaced -21.674±2.150 56.926±3.005 -84.115±2.462 0.932±0.058 1.678±0.164 0.983±0.258
S17 Non-Replaced -10.453±2.124 54.767±3.753 -76.125±2.259 0.436±0.034 1.096±0.061 0.699±0.169
S19 Non-Replaced -10.210±1.974 53.299±4.063 -74.121±0.465 0.686±0.068 1.065±0.070 1.157±0.102
S29 Non-Replaced -11.222±1.674 54.374±3.146 -73.938±1.861 0.663±0.082 1.100±0.027 0.310±0.061
S36 Non-Replaced -19.918±1.411 60.095±2.064 -85.515±0.993 0.727±0.054 0.872±0.052 0.734±0.120
Mean Replaced -12.7±5.0 55.3±5.3 -78.8±5.7 0.55±0.20 1.08±0.26 0.74±0.23
Mean Non-Replaced -17.0±5.1 56.3±4.5 -80.5±5.9 0.70±0.24 1.23±0.23 0.83±0.29
231
Table 30. Knee Flexion ROM (°) and loading-response knee extension moment (Nm/kg) for healthy controls. Flexion ROM Loading-response Extension Moment
Subject Limb Walking Ascent Descent Walking Ascent Descent
S5 Dominant -13.157±0.768 49.566±1.071 -66.403±1.468 1.085±0.061 1.693±0.068 1.350±0.095
S15 Dominant -20.972±1.611 55.796±2.821 -78.230±1.427 0.884±0.032 1.400±0.068 1.299±0.215
S21 Dominant -22.280±2.743 61.178±2.205 -85.547±1.369 0.946±0.063 1.535±0.065 1.393±0.108
S22 Dominant -18.821±1.704 52.666±0.662 -78.885±1.950 0.887±0.053 1.433±0.051 0.908±0.145
S23 Dominant -24.237±1.341 58.677±1.717 -80.497±1.770 0.857±0.084 1.281±0.109 0.580±0.205
S24 Dominant -17.712±2.156 54.717±1.347 -81.838±1.719 0.642±0.046 1.350±0.031 0.576±0.150
S25 Dominant -19.330±2.113 57.313±1.726 -76.734±0.874 0.855±0.095 0.897±0.034 0.682±0.065
S26 Dominant -22.473±1.888 58.529±1.685 -84.350±1.134 0.787±0.064 1.116±0.048 0.803±0.069
S27 Dominant -18.078±1.032 58.058±1.786 -86.126±1.331 0.640±0.053 1.329±0.035 0.848±0.078
S31 Dominant -20.228±1.792 55.363±0.587 -82.397±1.788 0.591±0.036 1.296±0.046 0.758±0.193
S33 Dominant -20.333±1.385 59.096±1.422 -79.063±1.357 0.895±0.048 1.175±0.063 1.011±0.089
S34 Dominant -17.034±0.720 54.272±3.821 -74.355±2.085 0.624±0.043 1.015±0.067 0.805±0.115
S37 Dominant -14.461±2.056 62.159±1.278 -86.826±4.646 0.408±0.043 1.100±0.061 0.579±0.266
S38 Dominant -12.054±1.682 67.537±1.903 -88.076±2.817 0.713±0.057 0.928±0.092 0.781±0.252
S39 Dominant -9.830±1.288 54.883±2.054 -82.451±3.755 0.249±0.006 0.863±0.063 0.311±0.089
S5 Non- Dominant -17.461±1.175 48.711±0.658 -73.302±0.927 1.111±0.078 1.817±0.051 1.089±0.151
S15 Non- Dominant -23.921±0.787 53.228±2.057 -78.838±2.731 0.941±0.087 1.163±0.074 1.240±0.150
S21 Non- Dominant -19.059±2.463 62.149±1.541 -87.911±1.576 0.890±0.047 1.428±0.063 1.043±0.061
S22 Non- Dominant -18.443±2.470 49.225±0.570 -78.117±1.663 0.922±0.037 1.382±0.037 0.639±0.179
S23 Non- Dominant -27.551±1.223 60.648±1.386 -76.072±2.041 0.838±0.034 1.509±0.114 0.749±0.177
S24 Non- Dominant -13.078±0.864 55.960±1.478 -81.712±1.230 0.678±0.014 1.294±0.031 0.747±0.099
S25 Non- Dominant -17.923±1.306 59.208±1.137 -71.956±2.117 0.744±0.034 1.009±0.069 0.581±0.127
S26 Non- Dominant -16.370±0.961 60.454±2.041 -86.449±1.529 0.548±0.041 0.984±0.023 0.745±0.057
S27 Non- Dominant -18.453±1.605 63.649±3.047 -86.532±3.027 0.968±0.068 1.417±0.036 0.750±0.114
S31 Non- Dominant -19.473±1.470 53.495±1.757 -86.605±1.264 0.619±0.017 0.887±0.062 0.435±0.192
S33 Non- Dominant -20.511±0.998 60.147±2.855 -79.057±1.499 0.776±0.032 1.081±0.051 0.778±0.149
S34 Non- Dominant -15.868±1.266 52.726±1.132 -69.314±2.535 0.514±0.095 0.974±0.043 0.569±0.045
S37 Non- Dominant -14.899±1.870 59.972±2.094 -82.525±1.941 0.551±0.025 1.004±0.034 0.720±0.117
S38 Non- Dominant -13.420±2.884 61.873±4.320 -80.080±3.014 0.693±0.035 0.880±0.110 0.531±0.104
S39 Non- Dominant -10.738±1.391 55.803±2.334 -78.919±2.086 0.270±0.052 0.982±0.089 0.207±0.110
Mean Dominant -18.1±4.1 57.3±4.3 -80.8±5.6 0.74±0.22 1.23±0.24 0.85±0.31
Mean Non- Dominant -17.8±4.2 57.1±4.8 -79.8±5.6 0.74±0.22 1.19±0.27 0.72±0.26
232
Table 31. Push-off knee extension moment (Nm/kg) and knee adduction ROM (°) for Dissatisfied TKR patients. Push-off Extension Moment Ab/Adduction ROM
Subject Limb Walking Ascent Descent Walking Ascent Descent
S4 Replaced 0.204±0.022 - 0.882±0.048 4.698±0.714 -17.550±1.558 11.965±1.954
S6 Replaced 0.271±0.027 - 0.826±0.031 2.708±0.339 -2.590±0.384 5.767±0.670
S13 Replaced 0.255±0.017 - 0.847±0.075 2.449±0.307 -5.492±1.303 13.946±0.755
S18 Replaced 0.196±0.014 - 1.126±0.082 0.229±0.053 -12.326±0.577 4.865±0.682
S20 Replaced 0.172±0.019 - 0.565±0.037 3.875±0.279 -18.083±1.474 9.936±2.175
S28 Replaced 0.106±0.038 - .±. 1.134±0.146 -7.533±1.044 4.611±0.744
S30 Replaced 0.074±0.011 - 0.671±0.079 2.950±0.352 -2.371±0.893 8.128±0.688
S32 Replaced 0.047±0.019 - 0.698±0.012 3.803±0.448 -20.472±0.870 15.808±1.225
S35 Replaced 0.085±0.015 - .±. 1.474±0.416 -19.666±1.045 26.218±1.007
S4 Non-Replaced 0.318±0.029 - 0.837±0.061 2.920±0.531 -14.458±2.734 4.549±1.756
S6 Non-Replaced 0.295±0.021 - 1.276±0.033 5.111±0.490 -23.779±0.468 12.695±0.829
S13 Non-Replaced 0.347±0.032 - 1.408±0.057 1.578±0.550 -7.290±1.237 12.691±1.168
S18 Non-Replaced 0.190±0.013 - 1.417±0.037 1.710±0.346 -16.059±1.151 8.844±0.337
S20 Non-Replaced 0.189±0.020 - 0.788±0.053 1.850±0.403 -10.302±1.816 3.824±0.928
S28 Non-Replaced 0.201±0.034 - 1.276±0.093 2.283±0.488 -7.890±1.963 11.465±1.227
S30 Non-Replaced 0.106±0.007 - 0.919±0.065 2.124±0.530 -1.220±0.637 6.801±0.978
S32 Non-Replaced 0.028±0.000 - 0.756±0.027 2.478±0.787 -17.884±1.654 10.062±3.514
S35 Non-Replaced 0.170±0.024 - 1.175±0.048 4.111±0.644 -11.093±1.081 6.130±0.811
Mean Replaced 0.16±0.08 - 0.78±0.17 2.6±1.4 -11.8±7.4 11.2±6.9
Mean Non-Replaced 0.20±0.10 - 1.09±0.27 2.7±1.2 -12.2±6.7 8.6±3.4
233
Table 32. Push-off knee extension moment (Nm/kg) and knee adduction ROM (°) for Satisfied TKR patients. Push-off Extension Moment Ab/Adduction ROM
Subject Limb Walking Ascent Descent Walking Ascent Descent
S1 Replaced 0.213±0.028 - 0.967±0.072 1.956±0.957 -15.123±0.778 1.948±1.029
S2 Replaced 0.359±0.023 - 1.016±0.133 1.974±0.309 -9.633±0.735 7.271±1.005
S3 Replaced 0.299±0.012 - 0.728±0.084 9.459±1.060 -8.760±1.484 7.719±1.341
S7 Replaced 0.227±0.010 - .±. 1.601±0.227 -14.734±0.977 16.086±1.393
S8 Replaced 0.375±0.034 - 1.306±0.065 0.124±0.376 -10.503±0.547 6.062±1.257
S9 Replaced 0.260±0.014 - 1.082±0.059 3.953±0.788 -22.868±1.030 16.105±1.190
S10 Replaced 0.305±0.017 - 0.819±0.015 2.984±0.908 -10.812±1.145 14.004±1.255
S11 Replaced 0.231±0.017 - 0.881±0.069 3.222±0.530 -21.185±1.945 9.806±0.000
S12 Replaced 0.228±0.010 - 0.410±0.044 2.185±0.506 -23.726±1.386 18.465±0.215
S14 Replaced 0.353±0.018 - 1.134±0.072 -1.634±0.941 -17.455±1.188 3.414±1.135
S16 Replaced 0.392±0.017 - 1.258±0.056 0.925±0.208 -14.814±8.685 14.217±0.829
S17 Replaced 0.184±0.018 - 0.782±0.033 0.663±0.391 -16.041±1.554 9.923±1.077
S19 Replaced .±. - 0.747±0.053 2.715±0.311 -16.673±1.173 3.773±0.702
S29 Replaced 0.053±0.006 - 0.719±0.039 2.396±0.681 -13.868±1.347 15.912±1.534
S36 Replaced 0.179±0.027 - 0.530±0.065 2.038±0.318 -22.074±1.064 11.549±1.099
S1 Non-Replaced 0.229±0.014 - 1.142±0.101 0.578±0.394 -9.087±1.468 7.623±1.013
S2 Non-Replaced 0.432±0.015 - 1.155±0.059 1.391±0.222 -8.947±0.665 3.857±0.907
S3 Non-Replaced 0.286±0.021 - 0.758±0.111 2.587±0.702 -6.050±2.924 13.208±0.357
S7 Non-Replaced 0.269±0.012 - 1.243±0.057 -0.242±0.728 -11.249±1.930 9.350±2.548
S8 Non-Replaced 0.234±0.018 - 0.936±0.041 0.481±0.695 -16.807±1.058 10.452±0.838
S9 Non-Replaced 0.302±0.022 - 1.262±0.097 1.220±0.174 -10.462±0.808 5.115±1.252
S10 Non-Replaced 0.316±0.029 - 1.038±0.029 -0.744±0.390 -3.992±1.313 5.142±1.516
S11 Non-Replaced 0.179±0.009 - 0.924±0.056 -0.501±0.435 -0.974±0.547 4.877±3.197
S12 Non-Replaced 0.306±0.028 - 1.134±0.056 2.743±0.241 -13.384±1.449 6.834±2.238
S14 Non-Replaced 0.316±0.035 - 0.871±0.121 2.161±0.249 -18.102±0.449 5.427±1.176
S16 Non-Replaced 0.347±0.010 - 1.610±0.031 -1.361±1.311 -10.976±0.996 2.466±0.666
S17 Non-Replaced 0.169±0.014 - 0.748±0.098 1.014±0.567 -13.016±3.791 9.024±1.660
S19 Non-Replaced 0.209±0.012 - 1.163±0.096 2.075±0.576 -1.775±0.986 6.967±1.081
S29 Non-Replaced 0.124±0.008 - 0.858±0.041 2.410±0.665 -2.808±0.478 6.320±1.157
S36 Non-Replaced 0.249±0.025 - 0.935±0.052 -0.754±0.324 -12.109±0.657 5.758±0.375
Mean Replaced 0.26±0.09 - 0.88±0.26 2.3±2.4 -15.9±4.9 10.4±5.3
Mean Non-Replaced 0.26±0.08 - 1.05±0.23 0.9±1.4 -9.3±5.3 6.8±2.8
234
Table 33. Push-off knee extension moment (Nm/kg) and knee adduction ROM (°) for healthy controls. Push-off Extension Moment Ab/Adduction ROM
Subject Limb Walking Ascent Descent Walking Ascent Descent
S5 Dominant 0.508±0.049 - 1.157±0.113 1.731±0.079 -6.632±0.419 10.834±1.227
S15 Dominant 0.284±0.023 - 0.846±0.045 .±. -10.693±1.782 12.258±1.420
S21 Dominant 0.208±0.016 - 1.121±0.129 2.900±0.365 -21.304±0.842 6.721±1.945
S22 Dominant 0.298±0.014 - 1.327±0.081 3.486±0.546 -5.535±2.719 5.433±1.773
S23 Dominant 0.116±0.016 - 0.999±0.045 0.471±0.221 -18.447±1.111 7.468±1.094
S24 Dominant 0.214±0.015 - 0.981±0.050 5.649±0.883 -19.612±8.896 .±.
S25 Dominant 0.099±0.030 - 0.959±0.035 4.452±0.887 -16.893±4.555 5.451±1.112
S26 Dominant 0.124±0.029 - 0.798±0.048 2.078±0.248 -7.900±0.848 5.167±0.396
S27 Dominant 0.187±0.028 - 0.731±0.085 7.439±0.428 -36.092±1.143 1.636±1.138
S31 Dominant 0.126±0.012 - 1.162±0.096 4.564±0.508 -3.598±0.593 9.079±0.987
S33 Dominant 0.207±0.023 - 1.053±0.045 0.178±0.415 -7.389±0.668 9.703±0.776
S34 Dominant 0.095±0.008 - 0.754±0.097 3.004±0.468 -20.576±0.544 10.832±0.937
S37 Dominant 0.169±0.009 - 0.829±0.079 3.174±0.842 -0.720±1.598 12.219±0.957
S38 Dominant 0.102±0.024 - 0.789±0.031 3.321±0.441 -18.209±1.519 8.652±1.026
S39 Dominant 0.274±0.030 - 0.881±0.042 4.355±0.230 -14.889±3.424 13.543±1.914
S5 Non- Dominant 0.513±0.035 - 1.027±0.090 1.230±0.323 -3.537±1.035 4.201±1.339
S15 Non- Dominant 0.279±0.024 - 1.073±0.059 3.831±1.019 -4.837±0.319 10.022±1.376
S21 Non- Dominant 0.244±0.019 - 1.131±0.055 1.339±0.148 -18.072±0.919 4.070±1.772
S22 Non- Dominant 0.286±0.038 - 1.102±0.070 0.531±0.259 -3.461±1.996 6.465±1.464
S23 Non- Dominant 0.119±0.022 - 1.144±0.079 4.131±0.328 -22.472±0.671 7.411±0.991
S24 Non- Dominant 0.190±0.018 - 1.064±0.027 3.622±0.430 -20.602±0.723 .±.
S25 Non- Dominant 0.076±0.006 - 0.982±0.101 2.923±0.311 -18.356±0.734 7.423±0.477
S26 Non- Dominant 0.197±0.011 - 0.638±0.027 5.982±0.403 -14.759±7.493 12.456±2.112
S27 Non- Dominant 0.178±0.032 - 0.784±0.037 0.981±0.220 -28.405±0.851 7.071±1.113
S31 Non- Dominant 0.148±0.010 - 0.935±0.029 1.519±0.515 -3.923±0.767 8.013±1.170
S33 Non- Dominant 0.149±0.062 - 0.868±0.057 1.986±0.154 -6.807±1.515 9.911±1.423
S34 Non- Dominant .±. - 0.747±0.092 1.180±0.606 -16.960±0.662 6.935±1.062
S37 Non- Dominant 0.155±0.014 - 0.701±0.044 0.072±0.121 -6.688±0.710 4.428±1.021
S38 Non- Dominant 0.118±0.013 - 0.833±0.112 1.636±0.181 -11.562±1.527 10.275±1.215
S39 Non- Dominant 0.260±0.024 - 0.966±0.042 2.439±0.237 -7.269±0.901 7.648±0.520
Mean Dominant 0.20±0.11 - 0.96±0.18 3.3±1.9 -13.9±9.1 8.5±3.4
Mean Non- Dominant 0.21±0.11 - 0.93±0.16 2.2±1.6 -12.5±8.0 7.6±2.5
235
Table 34. Loading-response and push-off knee abduction moments (Nm/kg) for Dissatisfied TKR patients. Loading-response Abduction Moment Push-off Abduction Moment
Subject Limb Walking Ascent Descent Walking Ascent Descent
S4 Replaced -0.567±0.052 -0.323±0.040 -0.613±0.039 -0.425±0.034 -0.209±0.023 -0.546±0.038
S6 Replaced -0.606±0.036 -0.645±0.048 -0.538±0.037 -0.564±0.027 -0.668±0.052 -0.547±0.042
S13 Replaced -0.591±0.020 -0.295±0.021 -0.630±0.020 -0.363±0.020 -0.126±0.011 -0.580±0.012
S18 Replaced -0.531±0.036 -0.542±0.030 -0.328±0.026 -0.370±0.014 -0.559±0.041 -0.189±0.020
S20 Replaced -0.450±0.011 -0.211±0.042 -0.504±0.022 -0.250±0.015 -0.107±0.052 -0.445±0.073
S28 Replaced -0.145±0.049 -0.182±0.046 -0.176±0.052 -0.192±0.076 -0.274±0.036 -0.092±0.018
S30 Replaced -0.344±0.048 -0.384±0.036 -0.275±0.120 -0.140±0.020 -0.296±0.031 -0.094±0.067
S32 Replaced -0.302±0.010 -0.469±0.022 -0.225±0.030 -0.204±0.020 -0.376±0.029 -0.280±0.030
S35 Replaced -0.225±0.026 -0.205±0.025 -0.285±0.051 -0.084±0.038 .±. -0.343±0.024
S4 Non-Replaced -0.534±0.028 -0.623±0.128 -0.536±0.036 -0.560±0.037 -0.716±0.046 -0.358±0.049
S6 Non-Replaced -0.459±0.028 -0.591±0.032 -0.781±0.026 -0.416±0.046 -0.237±0.011 -0.677±0.040
S13 Non-Replaced -0.472±0.024 -0.455±0.045 -0.334±0.045 -0.280±0.056 -0.469±0.030 -0.334±0.052
S18 Non-Replaced -0.663±0.015 -0.469±0.008 -0.692±0.051 -0.422±0.018 -0.205±0.024 -0.697±0.023
S20 Non-Replaced -0.367±0.013 -0.515±0.019 -0.434±0.071 -0.257±0.008 -0.443±0.056 -0.302±0.046
S28 Non-Replaced -0.486±0.045 -0.433±0.024 -0.634±0.056 -0.311±0.050 -0.229±0.061 -0.681±0.039
S30 Non-Replaced -0.531±0.052 -0.189±0.036 -0.580±0.049 -0.250±0.017 -0.063±0.018 -0.341±0.069
S32 Non-Replaced -0.415±0.022 -0.363±0.029 -0.507±0.035 -0.278±0.028 -0.232±0.029 -0.608±0.026
S35 Non-Replaced -0.530±0.053 -0.655±0.069 -0.574±0.030 -0.348±0.011 -0.513±0.026 -0.349±0.019
Mean Replaced -0.42±0.17 -0.36±0.16 -0.40±0.17 -0.29±0.15 -0.33±0.20 -0.35±-0.19
Mean Non-Replaced -0.50±0.08 -0.48±0.14 -0.56±0.13 -0.35±0.10 -0.35±0.20 -0.48±0.18
236
Table 35. Loading-response and push-off knee abduction moments (Nm/kg) for Satisfied TKR patients. Loading-response Abduction Moment Push-off Abduction Moment
Subject Limb Walking Ascent Descent Walking Ascent Descent
S1 Replaced -0.558±0.009 -0.597±0.050 -0.513±0.023 -0.378±0.043 -0.527±0.050 -0.327±0.039
S2 Replaced -0.536±0.017 -0.627±0.055 -0.541±0.066 -0.469±0.028 -0.521±0.008 -0.428±0.061
S3 Replaced -0.779±0.040 -0.435±0.048 -0.937±0.058 -0.376±0.019 -0.414±0.090 -0.649±0.080
S7 Replaced -0.420±0.024 -0.151±0.038 -0.411±0.046 -0.163±0.030 -0.113±0.045 -0.343±0.062
S8 Replaced -0.301±0.014 -0.386±0.012 -0.404±0.038 -0.323±0.013 -0.464±0.042 -0.199±0.016
S9 Replaced -0.552±0.014 -0.443±0.027 -0.787±0.039 -0.356±0.025 -0.139±0.019 -0.608±0.033
S10 Replaced -0.511±0.027 -0.185±0.037 -0.688±0.063 -0.262±0.019 .±. -0.317±0.009
S11 Replaced -0.500±0.028 -0.374±0.041 -0.829±0.042 -0.317±0.027 .±. -0.617±0.066
S12 Replaced -0.443±0.017 -0.403±0.017 -0.804±0.045 -0.328±0.014 .±. -0.512±0.027
S14 Replaced -0.511±0.027 -0.179±0.032 -0.616±0.019 -0.318±0.044 -0.074±0.071 -0.464±0.024
S16 Replaced -0.559±0.027 -0.378±0.077 -0.687±0.093 -0.460±0.029 -0.248±0.044 -0.544±0.025
S17 Replaced -0.241±0.036 -0.350±0.025 -0.247±0.057 -0.228±0.026 -0.305±0.019 -0.118±0.042
S19 Replaced -0.341±0.013 -0.210±0.055 -0.291±0.041 -0.216±0.027 -0.289±0.028 -0.053±0.016
S29 Replaced -0.558±0.017 -0.360±0.021 -0.866±0.041 -0.388±0.028 -0.110±0.066 -0.499±0.027
S36 Replaced -0.506±0.026 -0.191±0.033 -0.539±0.040 -0.359±0.009 -0.041±0.018 -0.402±0.018
S1 Non-Replaced -0.687±0.051 -0.360±0.057 -0.930±0.023 -0.334±0.030 -0.230±0.016 -0.646±0.062
S2 Non-Replaced -0.531±0.016 -0.202±0.035 -0.599±0.032 -0.403±0.012 -0.119±0.008 -0.418±0.036
S3 Non-Replaced -0.391±0.029 -0.494±0.071 -0.473±0.068 -0.392±0.039 -0.738±0.040 -0.306±0.063
S7 Non-Replaced -0.368±0.051 -0.431±0.048 -0.251±0.091 -0.262±0.022 -0.270±0.018 -0.144±0.064
S8 Non-Replaced -0.545±0.030 -0.103±0.026 -0.653±0.039 -0.368±0.025 -0.211±0.022 -0.484±0.041
S9 Non-Replaced -0.507±0.023 -0.545±0.067 -0.535±0.043 -0.515±0.011 -0.676±0.035 -0.275±0.037
S10 Non-Replaced -0.159±0.025 -0.149±0.036 -0.046±0.064 -0.131±0.028 -0.093±0.023 .±.
S11 Non-Replaced -0.344±0.011 -0.528±0.024 -0.353±0.040 -0.377±0.037 -0.511±0.039 -0.272±0.031
S12 Non-Replaced -0.547±0.028 -0.587±0.027 -0.602±0.047 -0.407±0.026 -0.601±0.060 -0.280±0.036
S14 Non-Replaced -0.541±0.029 -0.603±0.024 -0.528±0.017 -0.379±0.018 -0.582±0.034 -0.368±0.039
S16 Non-Replaced -0.532±0.022 -0.493±0.050 -0.613±0.051 -0.497±0.015 -0.698±0.023 -0.316±0.035
S17 Non-Replaced -0.310±0.031 -0.231±0.022 -0.613±0.060 -0.113±0.024 .±. -0.356±0.077
S19 Non-Replaced -0.584±0.051 -0.283±0.034 -0.982±0.033 -0.356±0.025 -0.218±0.028 -0.567±0.045
S29 Non-Replaced -0.630±0.040 -0.624±0.013 -0.625±0.042 -0.603±0.016 -0.637±0.049 -0.356±0.050
S36 Non-Replaced -0.478±0.035 -0.432±0.014 -0.424±0.072 -0.517±0.021 -0.595±0.018 -0.197±0.047
Mean Replaced -0.48±0.13 -0.35±0.15 -0.61±0.21 -0.33±0.09 -0.27±0.18 -0.41±0.18
Mean Non-Replaced -0.48±0.14 -0.40±0.17 -0.55±0.23 -0.38±0.13 -0.44±0.24 -0.36±0.14
237
Table 36. Loading-response and push-off knee abduction moments (Nm/kg) for healthy controls. Loading-response Abduction Moment Push-off Abduction Moment
Subject Limb Walking Ascent Descent Walking Ascent Descent
S5 Dominant -0.370±0.029 -0.152±0.027 -0.540±0.034 -0.249±0.009 .±. -0.460±0.048
S15 Dominant -0.648±0.018 -0.357±0.048 -1.117±0.058 -0.562±0.019 -0.151±0.050 -0.502±0.033
S21 Dominant -0.596±0.018 -0.661±0.039 -0.727±0.017 -0.310±0.028 -0.104±0.014 -0.491±0.043
S22 Dominant -0.649±0.034 -0.206±0.023 -0.738±0.047 -0.287±0.022 -0.095±0.019 -0.515±0.031
S23 Dominant -0.357±0.016 -0.202±0.035 -0.598±0.045 -0.308±0.030 .±. -0.363±0.066
S24 Dominant -0.707±0.029 -0.656±0.035 -0.635±0.065 -0.487±0.008 .±. -0.667±0.035
S25 Dominant -0.599±0.045 -0.369±0.043 -0.611±0.053 -0.327±0.018 -0.167±0.062 -0.524±0.050
S26 Dominant -0.472±0.012 -0.431±0.057 -0.365±0.071 -0.380±0.017 -0.525±0.024 -0.259±0.040
S27 Dominant -0.614±0.044 -0.682±0.060 -0.737±0.035 -0.252±0.028 .±. -0.426±0.037
S31 Dominant -0.528±0.017 -0.359±0.032 -0.740±0.040 -0.220±0.008 .±. -0.539±0.037
S33 Dominant -0.323±0.017 -0.042±0.012 -0.504±0.014 -0.205±0.030 .±. -0.486±0.032
S34 Dominant -0.776±0.020 -0.601±0.025 -1.006±0.079 -0.288±0.026 -0.166±0.060 -0.625±0.055
S37 Dominant -0.575±0.045 -0.435±0.027 -0.794±0.063 -0.366±0.026 -0.127±0.014 -0.651±0.040
S38 Dominant -0.481±0.034 -0.269±0.037 -0.667±0.025 -0.325±0.011 -0.109±0.003 -0.498±0.064
S39 Dominant -0.599±0.023 -0.386±0.026 -0.680±0.025 -0.354±0.025 -0.158±0.033 -0.671±0.043
S5 Non- Dominant -0.440±0.036 -0.404±0.048 -0.273±0.024 -0.253±0.018 -0.326±0.031 -0.090±0.065
S15 Non- Dominant -0.626±0.034 -0.679±0.037 -0.595±0.082 -0.487±0.023 -0.701±0.059 -0.293±0.038
S21 Non- Dominant -0.285±0.005 -0.435±0.029 -0.308±0.023 -0.379±0.032 -0.266±0.014 -0.225±0.029
S22 Non- Dominant -0.380±0.021 -0.301±0.030 -0.308±0.025 -0.198±0.031 -0.383±0.041 -0.184±0.014
S23 Non- Dominant -0.386±0.010 -0.517±0.021 -0.400±0.030 -0.312±0.025 -0.296±0.029 -0.202±0.026
S24 Non- Dominant -0.614±0.022 -0.848±0.025 -0.515±0.022 -0.475±0.028 -0.530±0.012 -0.426±0.025
S25 Non- Dominant -0.426±0.029 -0.720±0.040 -0.488±0.046 -0.372±0.017 -0.647±0.018 -0.427±0.075
S26 Non- Dominant -0.615±0.019 -0.339±0.016 -0.639±0.026 -0.254±0.020 .±. -0.459±0.019
S27 Non- Dominant -0.274±0.016 -0.549±0.037 -0.286±0.032 -0.248±0.035 -0.279±0.018 -0.143±0.025
S31 Non- Dominant -0.238±0.014 -0.378±0.040 -0.223±0.049 -0.224±0.014 -0.285±0.040 -0.178±0.048
S33 Non- Dominant -0.352±0.032 -0.388±0.025 -0.148±0.029 -0.200±0.012 -0.285±0.025 -0.211±0.014
S34 Non- Dominant -0.302±0.032 -0.535±0.040 -0.340±0.027 -0.111±0.020 -0.355±0.028 -0.210±0.030
S37 Non- Dominant -0.453±0.025 -0.425±0.037 -0.294±0.084 -0.420±0.037 -0.520±0.061 -0.155±0.028
S38 Non- Dominant -0.425±0.027 -0.480±0.073 -0.288±0.033 -0.380±0.025 -0.585±0.077 -0.279±0.038
S39 Non- Dominant -0.503±0.030 -0.544±0.049 -0.489±0.036 -0.489±0.035 -0.633±0.014 -0.360±0.042
Mean Dominant -0.55±0.13 -0.39±0.20 -0.70±0.19 -0.33±0.10 -0.18±0.13 -0.51±0.11
Mean Non- Dominant -0.42±0.13 -0.50±0.15 -0.37±0.14 -0.32±0.12 -0.44±0.16 -0.26±0.11
238
Table 37. Ankle dorsiflexion ROM (°), loading-response dorsiflexion moment (Nm/kg), and push-off plantarflexion moment during
walking for Dissatisfied TKR patients. Subject Limb Dorsiflexion ROM Dorsiflexion Moment Plantarflexion Moment
S4 Replaced 25.506±1.201 0.393±0.022 -1.332±0.023
S6 Replaced 22.446±1.250 0.098±0.019 -1.279±0.053
S13 Replaced 27.109±1.143 0.374±0.014 -1.376±0.033
S18 Replaced 22.815±1.541 0.356±0.036 -1.261±0.035
S20 Replaced 25.562±0.663 0.214±0.015 -1.063±0.016
S28 Replaced 20.274±1.259 0.205±0.052 -1.206±0.024
S30 Replaced 22.953±0.908 0.281±0.029 -1.122±0.017
S32 Replaced 24.973±0.840 0.361±0.037 -1.077±0.033
S35 Replaced 23.329±1.353 0.263±0.019 -1.081±0.054
S4 Non-Replaced 21.976±2.316 0.259±0.013 -1.365±0.054
S6 Non-Replaced 23.263±0.335 0.236±0.031 -1.368±0.017
S13 Non-Replaced 13.916±0.879 0.309±0.017 -1.460±0.059
S18 Non-Replaced 15.463±2.438 0.224±0.014 -1.357±0.059
S20 Non-Replaced 22.950±1.590 0.247±0.020 -1.194±0.039
S28 Non-Replaced 24.089±1.316 0.095±0.011 -1.141±0.030
S30 Non-Replaced 19.923±0.382 0.334±0.022 -1.139±0.029
S32 Non-Replaced 21.541±1.284 0.320±0.032 -1.126±0.045
S35 Non-Replaced 25.193±0.422 0.229±0.023 -1.058±0.043
Mean Replaced 23.9±2.1 0.28±0.10 -1.20±0.12
Mean Non-Replaced 20.9±3.9 0.25±0.07 -1.25±0.14
239
Table 38. Ankle dorsiflexion ROM (°), loading-response dorsiflexion moment (Nm/kg), and push-off plantarflexion moment during
walking for Satisfied TKR patients. Subject Limb Dorsiflexion ROM Dorsiflexion Moment Plantarflexion Moment
S1 Replaced 28.185±0.913 0.487±0.029 -1.513±0.043
S2 Replaced 20.982±1.621 0.255±0.018 -1.480±0.012
S3 Replaced 24.835±1.274 0.455±0.013 -1.188±0.044
S7 Replaced 25.741±0.603 0.359±0.034 -1.202±0.022
S8 Replaced 30.579±1.625 0.176±0.010 -1.607±0.062
S9 Replaced 27.260±1.715 0.382±0.019 -1.487±0.048
S10 Replaced 25.445±1.470 0.264±0.015 -1.450±0.031
S11 Replaced 27.429±1.919 0.253±0.027 -1.132±0.045
S12 Replaced 21.537±0.770 0.264±0.023 -1.181±0.020
S14 Replaced 24.467±1.771 0.496±0.022 -1.457±0.047
S16 Replaced 23.481±1.371 0.295±0.029 -1.288±0.054
S17 Replaced 20.552±0.914 0.222±0.027 -1.452±0.040
S19 Replaced 28.419±0.874 0.399±0.017 -1.312±0.053
S29 Replaced 20.844±0.428 0.298±0.021 -1.434±0.024
S36 Replaced 23.743±1.734 0.428±0.029 -1.419±0.033
S1 Non-Replaced 20.186±3.596 0.475±0.047 -1.597±0.045
S2 Non-Replaced 19.705±1.235 0.253±0.018 -1.466±0.005
S3 Non-Replaced 20.160±0.711 0.387±0.022 -1.465±0.035
S7 Non-Replaced 20.137±0.986 0.293±0.028 -1.415±0.049
S8 Non-Replaced 32.213±0.764 0.311±0.014 -1.541±0.016
S9 Non-Replaced 28.591±1.224 0.292±0.011 -1.497±0.019
S10 Non-Replaced 23.975±0.535 0.239±0.020 -1.564±0.042
S11 Non-Replaced 22.949±0.727 0.195±0.030 -1.224±0.038
S12 Non-Replaced 19.443±1.022 0.277±0.024 -1.256±0.024
S14 Non-Replaced 24.695±0.780 0.393±0.027 -1.669±0.039
S16 Non-Replaced 25.352±1.362 0.249±0.016 -1.477±0.027
S17 Non-Replaced 15.935±1.842 0.192±0.023 -1.362±0.042
S19 Non-Replaced 20.988±0.844 0.370±0.033 -1.291±0.036
S29 Non-Replaced 19.421±1.076 0.306±0.035 -1.400±0.025
S36 Non-Replaced 17.154±1.235 0.404±0.031 -1.494±0.014
Mean Replaced 24.9±3.1 0.34±0.10 -1.37±0.14
Mean Non-Replaced 22.1±4.3 0.31±0.08 -1.45±0.13
240
Table 39. Ankle dorsiflexion ROM (°), loading-response dorsiflexion moment (Nm/kg), and push-off plantarflexion moment during
walking for healthy controls. Subject Limb Dorsiflexion ROM Dorsiflexion Moment Plantarflexion Moment
S5 Dominant 10.555±0.900 0.228±0.025 -1.165±0.045
S15 Dominant 19.703±1.558 0.273±0.022 -1.589±0.028
S21 Dominant 19.162±0.872 0.334±0.022 -1.334±0.032
S22 Dominant 17.174±1.197 0.476±0.029 -1.419±0.034
S23 Dominant 19.659±0.982 0.253±0.012 -1.553±0.033
S24 Dominant 20.261±2.013 0.390±0.043 -1.575±0.064
S25 Dominant 17.708±0.609 0.411±0.033 -1.517±0.011
S26 Dominant 21.248±0.901 0.392±0.026 -1.550±0.028
S27 Dominant 10.782±2.139 0.282±0.024 -1.265±0.022
S31 Dominant 25.798±0.324 0.169±0.015 -1.323±0.019
S33 Dominant 20.853±1.549 0.246±0.025 -1.327±0.025
S34 Dominant 16.575±1.482 0.402±0.038 -1.383±0.030
S37 Dominant 19.184±0.812 0.340±0.032 -1.276±0.048
S38 Dominant 19.731±1.355 0.388±0.023 -1.364±0.016
S39 Dominant 20.611±1.607 0.276±0.019 -1.209±0.072
S5 Non- Dominant 19.101±0.849 0.323±0.023 -1.542±0.022
S15 Non- Dominant 14.340±2.102 0.316±0.023 -1.490±0.050
S21 Non- Dominant 22.935±0.843 0.331±0.026 -1.342±0.042
S22 Non- Dominant 17.823±0.799 0.361±0.010 -1.624±0.019
S23 Non- Dominant 24.977±0.774 0.263±0.009 -1.444±0.045
S24 Non- Dominant 15.097±1.408 0.441±0.015 -1.529±0.042
S25 Non- Dominant 19.013±0.735 0.337±0.025 -1.552±0.033
S26 Non- Dominant 20.481±0.939 0.398±0.021 -1.345±0.027
S27 Non- Dominant 14.298±0.791 0.251±0.012 -1.419±0.045
S31 Non- Dominant 23.968±0.889 0.182±0.010 -1.323±0.020
S33 Non- Dominant 24.173±1.217 0.231±0.029 -1.253±0.038
S34 Non- Dominant 12.651±1.038 0.259±0.039 -1.425±0.041
S37 Non- Dominant 19.637±1.415 0.308±0.012 -1.328±0.044
S38 Non- Dominant 19.058±0.398 0.352±0.020 -1.426±0.039
S39 Non- Dominant 19.749±0.853 0.246±0.017 -1.325±0.015
Mean Dominant 18.6±3.9 0.32±0.09 -1.39±0.14
Mean Non- Dominant 19.2±3.8 0.31±0.07 -1.42±0.11
241
Table 40. Hip extension ROM (°), loading-response extension moment (Nm/kg), and push-off flexion moment (Nm/kg) during
walking for Dissatisfied TKR patients. Subject Limb Extension ROM Extension Moment Flexion Moment
S4 Replaced -40.775±0.854 -0.424±0.027 0.784±0.042
S6 Replaced -36.056±0.516 -0.398±0.046 0.602±0.054
S13 Replaced -44.977±0.831 -0.823±0.031 0.828±0.038
S18 Replaced -26.681±1.293 -0.349±0.025 0.449±0.043
S20 Replaced -32.993±0.269 -0.467±0.017 0.605±0.021
S28 Replaced -34.190±0.909 -0.325±0.023 0.258±0.216
S30 Replaced -32.821±1.049 -0.474±0.044 0.696±0.031
S32 Replaced -32.044±1.688 -0.533±0.022 0.322±0.045
S35 Replaced -31.981±1.350 -0.467±0.072 0.535±0.048
S4 Non-Replaced -43.479±2.476 -0.460±0.047 0.705±0.029
S6 Non-Replaced -37.235±1.057 -0.627±0.033 0.538±0.019
S13 Non-Replaced -46.345±1.148 -0.693±0.025 0.910±0.094
S18 Non-Replaced -43.780±0.688 -0.487±0.051 0.411±0.023
S20 Non-Replaced -34.384±0.680 -0.423±0.034 0.667±0.016
S28 Non-Replaced -25.167±0.543 -0.414±0.045 0.547±0.060
S30 Non-Replaced -27.676±0.447 -0.615±0.018 0.595±0.036
S32 Non-Replaced -34.097±1.363 -0.615±0.078 0.382±0.037
S35 Non-Replaced -27.178±1.376 -0.324±0.020 0.534±0.036
Mean Replaced -34.7±5.4 -0.47±0.15 0.56±0.20
Mean Non-Replaced -35.5±7.9 -0.52±0.12 0.59±0.16
242
Table 41. Hip extension ROM (°), loading-response extension moment (Nm/kg), and push-off flexion moment (Nm/kg) during
walking for Satisfied TKR patients. Subject Limb Extension ROM Extension Moment Flexion Moment
S1 Replaced -36.682±1.593 -0.964±0.089 0.774±0.033
S2 Replaced -28.065±0.742 -0.705±0.045 0.644±0.025
S3 Replaced -31.354±1.292 -0.976±0.062 0.549±0.031
S7 Replaced -36.614±0.825 -0.592±0.049 0.716±0.032
S8 Replaced -36.472±0.823 -0.636±0.038 0.714±0.044
S9 Replaced -37.271±1.093 -0.710±0.034 0.759±0.039
S10 Replaced -35.051±1.235 -0.658±0.040 0.582±0.647
S11 Replaced -36.591±1.569 -0.413±0.033 0.501±0.039
S12 Replaced -29.163±1.179 -0.580±0.027 0.517±0.030
S14 Replaced -28.915±0.770 -0.816±0.043 0.716±0.041
S16 Replaced -42.706±1.069 -0.893±0.078 0.561±0.041
S17 Replaced -37.476±0.597 -0.243±0.240 0.740±0.035
S19 Replaced -29.314±0.634 -0.605±0.040 0.355±0.046
S29 Replaced -30.734±0.444 -0.586±0.034 0.630±0.013
S36 Replaced -46.634±0.962 -0.622±0.036 0.961±0.043
S1 Non-Replaced -34.975±1.998 -0.867±0.068 0.604±0.063
S2 Non-Replaced -30.985±0.949 -0.662±0.034 0.688±0.025
S3 Non-Replaced -31.814±1.050 -1.223±0.071 0.419±0.034
S7 Non-Replaced -34.149±1.475 -0.577±0.045 0.709±0.033
S8 Non-Replaced -26.497±0.645 -0.854±0.028 0.433±0.037
S9 Non-Replaced -37.810±0.375 -0.598±0.046 0.665±0.037
S10 Non-Replaced -31.628±17.688 -0.583±0.081 1.027±0.085
S11 Non-Replaced -33.652±0.584 -0.304±0.023 0.508±0.027
S12 Non-Replaced -32.564±0.513 -0.530±0.040 0.580±0.027
S14 Non-Replaced -38.180±1.852 -0.818±0.031 0.676±0.081
S16 Non-Replaced -40.654±1.181 -0.842±0.053 0.573±0.034
S17 Non-Replaced -37.204±0.988 -0.359±0.045 0.839±0.034
S19 Non-Replaced -37.766±0.458 -0.343±0.041 0.650±0.016
S29 Non-Replaced -42.666±1.389 -0.630±0.097 0.510±0.045
S36 Non-Replaced -40.966±0.840 -0.606±0.049 0.904±0.042
Mean Replaced -34.9±5.3 -0.67±0.19 0.65±0.15
Mean Non-Replaced -35.4±4.4 -0.65±0.24 0.65±0.17
243
Table 42. Hip extension ROM (°), loading-response extension moment (Nm/kg), and push-off flexion moment (Nm/kg) during
walking for healthy controls. Subject Limb Extension ROM Extension Moment Flexion Moment
S5 Dominant -36.070±0.560 -0.763±0.016 0.465±0.027
S15 Dominant -38.848±1.422 -0.633±0.032 0.708±0.060
S21 Dominant -41.236±1.251 -0.435±0.050 0.771±0.020
S22 Dominant -38.244±0.537 -0.878±0.072 1.227±0.031
S23 Dominant -41.169±1.743 -0.930±0.027 0.609±0.060
S24 Dominant -33.716±1.535 -0.660±0.078 0.784±0.069
S25 Dominant -40.627±0.671 -0.425±0.026 0.733±0.064
S26 Dominant -43.972±1.266 -0.865±0.027 0.627±0.038
S27 Dominant -38.149±0.501 -0.368±0.044 0.833±0.046
S31 Dominant -30.848±0.669 -0.389±0.036 0.497±0.032
S33 Dominant -38.800±0.286 -0.592±0.049 0.544±0.015
S34 Dominant -37.998±1.020 -0.598±0.069 0.491±0.040
S37 Dominant -38.567±0.614 -0.682±0.063 0.788±0.074
S38 Dominant -43.462±0.814 -0.482±0.077 0.797±0.039
S39 Dominant -30.476±0.803 -0.470±0.027 0.671±0.032
S5 Non- Dominant -37.434±0.747 -0.461±0.021 0.713±0.048
S15 Non- Dominant -36.172±0.758 -0.540±0.032 0.788±0.030
S21 Non- Dominant -39.967±0.741 -0.444±0.033 0.796±0.054
S22 Non- Dominant -38.625±0.767 -0.970±0.061 1.165±0.095
S23 Non- Dominant -37.564±1.547 -0.755±0.063 0.727±0.047
S24 Non- Dominant -46.536±0.085 -0.513±0.037 0.724±0.030
S25 Non- Dominant -39.372±1.428 -0.516±0.045 0.589±0.024
S26 Non- Dominant -26.160±0.613 -0.779±0.026 0.644±0.019
S27 Non- Dominant -38.100±0.389 -0.276±0.068 0.760±0.053
S31 Non- Dominant -38.635±1.047 -0.365±0.022 0.458±0.023
S33 Non- Dominant -32.835±0.829 -0.759±0.057 0.407±0.043
S34 Non- Dominant -31.587±0.902 -0.513±0.045 0.361±0.040
S37 Non- Dominant -43.953±0.595 -0.591±0.052 0.737±0.031
S38 Non- Dominant -40.819±1.711 -0.370±0.073 0.717±0.038
S39 Non- Dominant -41.611±0.919 -0.450±0.022 0.598±0.029
Mean Dominant -38.1±4.0 -0.61±0.19 0.70±0.19
Mean Non- Dominant -38.0±5.0 -0.55±0.19 0.68±0.19
244
Table 43. Peak knee extension and flexion torque (Nm) for Dissatisfied TKR patients. Peak Extension Torque Peak Flexion Torque
Subject Limb 60°/s 180°/s 60°/s 180°/s
S4 Replaced 94.5 63.6 49.2 34.4
S6 Replaced 119.5 100.5 66.6 41.4
S13 Replaced 72.8 49.4 53.7 42.6
S18 Replaced 112.7 67.3 67.9 43.3
S20 Replaced 77.4 56.3 49 43.4
S28 Replaced 49.4 26.7 17.6 3.5
S30 Replaced 49.5 39.5 32.8 28.2
S32 Replaced 131.7 80.4 54.5 47.1
S35 Replaced 59.5 41 41 25.5
S4 Non-Replaced 96 66.9 35.5 26.8
S6 Non-Replaced 173 105.4 75 45.6
S13 Non-Replaced 120.8 81.9 66.2 44.7
S18 Non-Replaced 144.8 84.3 66.2 36.3
S20 Non-Replaced 91.3 71.5 53.7 31.9
S28 Non-Replaced 49.8 27.7 20.9 6.9
S30 Non-Replaced 66.7 44.1 33 28.1
S32 Non-Replaced 154.7 111.5 75.5 59.8
S35 Non-Replaced 97.8 62.5 45.7 22.8
Mean Replaced 85.2±30.9 58.3±22.7 48.0±15.9 34.4±13.7
Mean Non-Replaced 110.5±41.1 72.9±26.9 52.4±19.8 33.7±15.3
245
Table 44. Peak knee extension and flexion torque (Nm) for Satisfied TKR patients. Peak Extension Torque Peak Flexion Torque
Subject Limb 60°/s 180°/s 60°/s 180°/s
S1 Replaced 139.1 88 68.9 47.5
S2 Replaced 177.8 114.2 110 83
S3 Replaced 109.7 67.4 66.7 67
S7 Replaced 57.9 45 35.3 20.9
S8 Replaced 144.3 88.4 69.2 49.1
S9 Replaced 159.3 106.3 71.2 62.5
S10 Replaced 66.9 47.1 27 18.6
S11 Replaced 73.5 67.5 36.1 30.2
S12 Replaced 75.5 56.4 61.6 37.6
S14 Replaced 131.8 102.1 68.9 52.5
S16 Replaced 174.4 124.3 75.1 60.5
S17 Replaced 96.3 62.4 46.6 30.6
S19 Replaced 133.2 103.5 75.9 74.7
S29 Replaced 137.1 98.2 84.1 43.1
S36 Replaced 97.4 74.6 68.5 49.6
S1 Non-Replaced 186.6 140.8 79.9 61.7
S2 Non-Replaced 186 128.3 104.3 102.9
S3 Non-Replaced 135.7 77.3 61.7 66.3
S7 Non-Replaced 79.5 60.9 39.1 27.7
S8 Non-Replaced 107 99.7 84.9 72.8
S9 Non-Replaced 169.5 111.7 74.9 55.2
S10 Non-Replaced 62.8 54.5 24.5 22.4
S11 Non-Replaced 98.7 87.9 52.7 46.5
S12 Non-Replaced 126.4 74.3 65.6 57.8
S14 Non-Replaced 141.2 107.7 81.1 67.1
S16 Non-Replaced 201.5 131.9 85.6 64.4
S17 Non-Replaced 101.2 57.8 41.9 25.9
S19 Non-Replaced 128.8 105.6 73 81.5
S29 Non-Replaced 156.8 101.6 78.2 49.9
S36 Non-Replaced 105 90.4 59.4 52.6
Mean Replaced 118.3±39.1 83.0±24.9 64.3±21.1 48.5±19.0
Mean Non-Replaced 132.5±41.3 95.4±27.0 67.1±21.0 57.0±21.4
246
Table 45. Peak knee extension and flexion torque (Nm) for healthy controls. Peak Extension Torque Peak Flexion Torque
Subject Limb 60°/s 180°/s 60°/s 180°/s
S5 Dominant 109.7 78.4 65.2 57.5
S15 Dominant 176.1 118.8 84.1 61
S21 Dominant 158.7 116.2 70 56.1
S22 Dominant 102.4 71.2 59.9 42.2
S23 Dominant 131.1 88.4 62.1 38.1
S24 Dominant 147.5 97.1 76.3 57.2
S25 Dominant 154.3 84.2 82.9 57.9
S26 Dominant 99.1 77.8 73.9 64.1
S27 Dominant 161.6 131 65.4 54.5
S31 Dominant 54.9 35.8 41.6 25.1
S33 Dominant 126.8 81 55.6 40.8
S34 Dominant 91.1 97.1 62.1 69.6
S37 Dominant 96 47.7 55.7 26
S38 Dominant 79.7 55.1 52.2 39.3
S39 Dominant 75.4 32.8 41.6 25.5
S5 Non- Dominant 121.2 84.9 49.8 58.2
S15 Non- Dominant 159.5 105.4 96.5 54.4
S21 Non- Dominant 123.7 87.2 52.5 44.3
S22 Non- Dominant 112.1 73.5 57.9 38
S23 Non- Dominant 165.6 81.9 62 33.8
S24 Non- Dominant 139.1 93.6 79.5 67.5
S25 Non- Dominant 177.2 128.3 75.1 56.8
S26 Non- Dominant 89.4 67.5 60.6 57.1
S27 Non- Dominant 174.2 82 62.1 49.9
S31 Non- Dominant 45.4 21.3 26.7 33.9
S33 Non- Dominant 136.3 80.3 51 40.8
S34 Non- Dominant 137.8 87.6 64 58.3
S37 Non- Dominant 95.5 48.7 49.8 31.3
S38 Non- Dominant 76.7 54.8 40.4 38.9
S39 Non- Dominant 77.6 49.4 42 23.1
Mean Dominant 117.6±36.4 80.8±29.3 63.2±13.0 47.7±14.8
Mean Non- Dominant 122.1±39.3 76.4±25.8 58.0±17.1 45.8±12.7
247
Table 46. Unilateral overall stability index (OSI), mediolateral stability index (MLSI) and anteroposterior stability index (APSI) for
Dissatisfied TKR patients. Static Unilateral Dynamic Unilateral
Subject Limb OSI APSI MLSI OSI APSI MLSI
S4 Replaced 3.6 1.8 2.9 . . .
S6 Replaced 1.4 1.1 0.6 . . .
S13 Replaced 3.1 0.8 3 2.5 0.9 2.1
S18 Replaced 1.2 0.7 0.8 1.4 0.8 1
S20 Replaced . . . . . .
S28 Replaced 4.1 3.1 2.5 2.7 2.7 0.2
S30 Replaced 2.7 1.8 1.9 2.3 2.2 0.3
S32 Replaced . . . . . .
S35 Replaced . . . . . .
S4 Non-Replaced 2.3 1.7 1.3 . . .
S6 Non-Replaced 1.9 1.2 1.3 . . .
S13 Non-Replaced 2.5 2 1.2 3.1 2.7 1.2
S18 Non-Replaced . . . . . .
S20 Non-Replaced 5.1 4.6 1.7 2.8 2.5 1.1
S28 Non-Replaced 2.7 2.1 1.6 2.5 2.5 0.6
S30 Non-Replaced 4.7 2.8 3.7 2 1.8 0.8
S32 Non-Replaced . . . . . .
S35 Non-Replaced . . . . . .
Mean Replaced 2.68±1.17 1.55±0.90 1.95±1.04 2.23±0.57 1.65±0.95 0.90±0.88
Mean Non-Replaced 3.20±1.35 2.40±1.20 1.80±0.95 2.60±0.47 2.38±0.39 0.93±0.28
248
Table 47. Unilateral overall stability index (OSI), mediolateral stability index (MLSI) and anteroposterior stability index (APSI) for
Satisfied TKR patients. Static Unilateral Dynamic Unilateral
Subject Limb OSI APSI MLSI OSI APSI MLSI
S1 Replaced 1.4 0.6 1.1 . . .
S2 Replaced 1.6 1.3 0.7 4.4 0.7 4.3
S3 Replaced 6.3 5.2 3.4 2.7 2.6 0.4
S7 Replaced 3.3 0.7 3.1 2 1.4 1.3
S8 Replaced 1.7 1 1.2 2.4 1.9 1.2
S9 Replaced 4 0.7 3.9 3.1 0.8 2.9
S10 Replaced 5.3 3.3 4 1.8 1.1 1.1
S11 Replaced . . . . . .
S12 Replaced 3.3 1.9 2.5 2.7 1.5 2
S14 Replaced 4.4 3 2.9 3 2.3 1.9
S16 Replaced . . . . . .
S17 Replaced 2.4 1.8 1.2 2.5 2.1 1.1
S19 Replaced 1 0.7 0.6 2 1.4 1.2
S29 Replaced 2.7 1 2.3 . . .
S36 Replaced 2.3 1.2 1.7 1.4 0.2 0.6
S1 Non-Replaced . . . . . .
S2 Non-Replaced 5.7 4.8 2.9 . . .
S3 Non-Replaced 1.3 1.1 0.6 2.4 1.6 1.8
S7 Non-Replaced 1.4 0.6 1.1 2.5 2.3 0.6
S8 Non-Replaced 1.3 0.5 1.1 3.6 3.1 1.4
S9 Non-Replaced 1.3 0.9 0.8 4.4 0.7 4.3
S10 Non-Replaced 1.6 1.2 0.8 1.8 0.9 1.4
S11 Non-Replaced . . . . . .
S12 Non-Replaced 2.3 2.1 0.6 3 2.8 0.9
S14 Non-Replaced 1.1 0.7 0.6 2.1 2 0.5
S16 Non-Replaced . . . . . .
S17 Non-Replaced 4.6 1.3 4.3 3.2 1.4 2.8
S19 Non-Replaced 2.5 1.3 1.8 2 1.3 1.1
S29 Non-Replaced 1.5 1.2 0.6 2.1 1.1 1.6
S36 Non-Replaced 1.9 0.8 1.5 1.1 0.5 0.1
Mean Replaced 3.05±1.59 1.72±1.36 2.20±1.20 2.55±0.81 1.45±0.73 1.64±1.12
Mean Non-Replaced 2.21±1.46 1.38±1.16 1.39±1.14 2.56±0.92 1.61±0.85 1.50±1.18
249
Table 48. Unilateral overall stability index (OSI), mediolateral stability index (MLSI) and anteroposterior stability index (APSI) for
healthy controls. Static Unilateral Dynamic Unilateral
Subject Limb OSI APSI MLSI OSI APSI MLSI
S5 Dominant 3.5 1 3.3 2.1 1.3 1.4
S15 Dominant 3.8 2.2 3 2.5 1.4 2
S21 Dominant 4.4 0.9 4.2 2.3 1.1 1.8
S22 Dominant 5 3 3.8 1.8 1.3 1.1
S23 Dominant 2 0.7 1.7 2.4 0.7 2.2
S24 Dominant 3.2 1.2 2.9 2.1 1.5 1.2
S25 Dominant 2.2 0.5 2 1.7 0.8 1.4
S26 Dominant 2.2 0.7 2 3.5 3.1 1.3
S27 Dominant 1.1 0.4 0.9 1.8 1.3 0.9
S31 Dominant 6.2 5.2 3 2.3 2 0.8
S33 Dominant 1.1 0.6 0.7 2.5 2.2 1.1
S34 Dominant 3.4 1.8 2.7 2.9 2.9 0.4
S37 Dominant 1.9 1.4 0.1 . . .
S38 Dominant 1.6 1 1 3 2.9 0.8
S39 Dominant 4.6 1.9 4 . . .
S5 Non- Dominant 1.9 1 1.5 2.8 0.5 2.7
S15 Non- Dominant 1.1 0.6 0.7 1.4 0.8 1
S21 Non- Dominant 1.6 1.2 0.9 3 0.7 2.9
S22 Non- Dominant 1.4 1.1 0.6 3.3 1.8 2.6
S23 Non- Dominant 0.8 0.5 0.5 1.5 0.9 1
S24 Non- Dominant 2.4 2 1 1.8 1.2 0.8
S25 Non- Dominant 0.8 0.4 0.6 0.9 0.6 0.6
S26 Non- Dominant 2 0.7 1.8 3.5 3.3 0.8
S27 Non- Dominant 0.9 0.6 0.5 1.4 1 0.8
S31 Non- Dominant 3.5 2.3 2.4 2.3 2.2 0.6
S33 Non- Dominant 0.9 0.6 0.5 2.7 1.9 1.8
S34 Non- Dominant 2.2 1.6 1.3 3.5 2.5 2.5
S37 Non- Dominant 1.9 1.4 1 2.8 2.1 1.5
S38 Non- Dominant 1.3 0.9 0.8 2.3 1.4 1.6
S39 Non- Dominant 2.9 2.2 1.3 3 1.4 2.4
Mean Dominant 3.08±1.52 1.50±1.26 2.35±1.28 2.38±0.52 1.73±0.81 1.27±0.51
Mean Non- Dominant 1.71±0.81 1.14±0.63 1.03±0.55 2.41±0.84 1.49±0.80 1.57±0.85
250
Table 49. Bilateral static and dynamic overall stability index (OSI), mediolateral stability index (MLSI) and anteroposterior stability index (APSI) for controls and TKR patients.
Subject Group Static OSI Static APSI Static MLSI Dynamic OSI Dynamic APSI Dynamic MLSI
S4 Dissatisfied 1 0.8 0.4 1.2 0.9 0.7
S6 Dissatisfied 0.6 0.4 0.2 1.4 1 0.7
S13 Dissatisfied 0.3 0.2 0.2 1.6 1 1.2
S18 Dissatisfied 0.4 0.3 0.2 0.9 0.6 0.6
S20 Dissatisfied 6.1 6.1 0.1 1.5 0.9 1
S28 Dissatisfied 0.4 0.3 0.2 2.2 1.9 0.8
S30 Dissatisfied 2.1 1.3 1.5 2.6 1.8 1.7
S32 Dissatisfied 0.7 0.4 0.4 1.7 1 1.2
S35 Dissatisfied 0.6 0.4 0.3 2.1 2.1 0.2
S5 Healthy 0.6 0.3 0.4 1.7 1.3 1
S15 Healthy 1 0.8 0.5 1.7 0.7 1.5
S21 Healthy 0.4 0.3 0.1 1.5 0.9 0.9
S22 Healthy 3 2.5 1.5 2.7 2.5 0.6
S23 Healthy 0.4 0.3 0.2 1.4 0.8 1
S24 Healthy 2.5 1.8 1.6 2.3 2.2 0.8
S25 Healthy 0.3 0.1 0.2 0.5 0.3 0.4
S26 Healthy 1.2 1.1 0.3 3 2.8 1
S27 Healthy 0.3 0.3 0.2 1.1 0.7 0.6
S31 Healthy 4.6 4.6 0.5 1.6 1.2 0.6
S33 Healthy 0.4 0.3 0.1 1.3 1 0.6
S34 Healthy 1.6 1.6 0.4 2.5 2.2 0.9
S37 Healthy 0.8 0.6 0.4 1.5 1.1 0.8
S38 Healthy 0.4 0.3 0.2 1.9 1.3 1.1
S39 Healthy 3.5 1 3.2 2.1 1.8 0.9
S1 Satisfied 0.2 0.1 0.1 1.3 0.9 0.8
S2 Satisfied 2.7 2.7 0.2 2.2 1 1.8
S3 Satisfied 0.4 0.3 0.2 0.9 0.7 0.4
S7 Satisfied 0.5 0.3 0.3 0.7 0.4 0.4
S8 Satisfied 1.9 1.2 1.4 1.3 1.1 0.5
S9 Satisfied 0.5 0.4 0.2 1.2 0.7 0.9
S10 Satisfied 1.2 0.4 1 1.1 0.8 0.6
S11 Satisfied 0.3 0.3 0.1 1 0.5 0.7
S12 Satisfied 2.8 2.7 0.2 2.9 2.8 0.4
S14 Satisfied 0.5 0.4 0.2 3.3 3 1
S16 Satisfied 0.6 0.4 0.3 1.3 1.1 0.5
S17 Satisfied 0.6 0.5 0.3 1.2 0.9 0.7
S19 Satisfied 0.5 0.3 0.3 1.3 1.1 0.6
S29 Satisfied 0.4 0.3 0.2 1.5 1 0.9
S36 Satisfied 0.4 0.2 0.2 2.2 2.1 0.7
Mean Dissatisfied 1.36±1.86 1.13±1.89 0.39±0.43 1.69±0.53 1.24±0.54 0.90±0.43
Mean Satisfied 0.90±0.86 0.70±0.85 0.35±0.36 1.56±0.75 1.21±0.79 0.73±0.35
Mean Healthy 1.40±1.36 1.06±1.20 0.65±0.84 1.79±0.65 1.39±0.74 0.85±0.27
251
VITA
Kevin Alan Valenzuela was the 3rd of four boys, born in 1985 to Penelope and John
Valenzuela. He graduated from Orange Lutheran High School in 2003. After high school, he
attended California State University Fullerton where he graduated with a B.A. in history and a
secondary education teaching credential. In 2012, he returned to California State University
Fullerton to pursue an M.S. in Kinesiology. Kevin completed his education at the University of
Tennessee, earning a PhD in Biomechanics in 2017. He has accepted a tenure track, assistant
professor position at California Polytechnic University in Pomona, CA.